[OSS RADAR] September 2026

[github-actions]

[OSS RADAR] September 2026

One rolling monthly thread for PlotPickle OSS Radar daily discovery reports.

Reports are deterministic discovery evidence for Human review. Surfaced repositories are not automatically adopted, installed, copied into PlotPickle, turned into curriculum, or converted into implementation work.

Human review remains required before any product, curriculum, architecture or dependency decision.

[github-actions]

OSS Radar — 2026-09-13

GitHub discovery

• Enabled lanes: 6
• Search queries: 19
• Per-query result cap: 10
• Theoretical maximum raw pointers: 190
• Raw repository pointers returned: 27
• Unique repositories examined: 27
• Retained after hard rejection: 27
• Hard rejected: 0
• Hard rejection reasons: none

Metadata discovery only: this stage follows GitHub repository pointers and metadata; it does not clone or inspect repository source trees.

Repository creation age — unique repositories examined

• Created today: 0
• Created yesterday: 0
• Created earlier: 27
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 27
• Seen by Radar yesterday: 0
• Seen by Radar before yesterday: 0

Top 5 Radar history

• 1. IdaAkiwumi/narrative-soundstage — first seen by Radar today; first seen 2026-09-13
• 2. samihalawa/visual-ui-debug-agent-mcp — first seen by Radar today; first seen 2026-09-13
• 3. VAIXLNS/VAIXLNS — first seen by Radar today; first seen 2026-09-13
• 4. test-zeus-ai/testzeus-hercules — first seen by Radar today; first seen 2026-09-13
• 5. Orkas-AI/Orkas — first seen by Radar today; first seen 2026-09-13

Top repositories for review: 5/5
Meets qualification threshold: 0/5
Below threshold but included for review: 5/5

The daily review queue contains the Top 5 real repositories from today's retained candidates. Qualification status is context only and never blanks the report.

Phase 3 classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Top Repositories for Review

1. IdaAkiwumi/narrative-soundstage — WATCH

A FREE script-to-performance engine for screenwriters & producers. Automates table reads with dynamic casting, high-fidelity TTS, and a real-time prompter. Built for film pre-production, game design, and narrative logic testing

• Review status: Below qualification threshold by 6.53 points; included because it ranked in today's Top 5.
• Score: 48.47/100
• Why PlotPickle should care: LEARN / Writers Craft.
• What can PlotPickle learn from this? Investigate how the project approaches design, narrative, screenwriting, storytelling; repository metadata alone does not prove implementation fit.
• PlotPickle target: LEARN / Writers Craft
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: UNKNOWN (unknown)
• Activity: last activity about 138 days ago
• Review history: first Radar review
• Score evidence: activity:recent, fork:no, lane:writer-craft, license:unknown

2. samihalawa/visual-ui-debug-agent-mcp — WATCH

Visual UI testing and debugging MCP server with Playwright, screenshots, DOM inspection, workflow validation, and visual diffs.

• Review status: Below qualification threshold by 17.12 points; included because it ranked in today's Top 5.
• Score: 47.88/100
• Why PlotPickle should care: Validation / Local Platform / Documentation.
• What can PlotPickle learn from this? Investigate how the project approaches automation, browser, testing, visual; repository metadata alone does not prove implementation fit.
• PlotPickle target: Validation / Local Platform / Documentation
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: ISC (known-open-source)
• Activity: last activity about 34 days ago
• Review history: first Radar review
• Score evidence: activity:recent, fork:no, lane:platform-engineering, license:known-open-source

3. VAIXLNS/VAIXLNS — WATCH

╔════════════════════════════════════أرى أن هذا أصبح أقرب إلى Blueprint سيادي وليس مجرد مخطط نظام. ومع ذلك، إذا كان الهدف أن يكون هذا الرسم المرجعي النهائي (Master Architecture Blueprint) فلا تزال هناك فجوة معمارية واحدة فقط تمنع اكتماله بنسبة عالية: ما زلت تمثل الطبقات، لكن لا تمثل “الأبعاد” التي تعبر جميع الطبقات. أي أن الرسم الحالي يجيب على سؤال “ماذا يوجد؟” لكنه لا يجيب بالكامل على “ما الذي يمر عبر الجميع؟”. إذا أردت أن يبقى صالحًا لعقود، فأقترح تحويله من 2D Layered Architecture إلى Multi-Dimensional Sovereign Architecture. ⸻ Ω Universal Cross-Layer Fabric هذه ليست طبقة جديدة، بل “نسيج” يخترق جميع الحضارات من Ω∞ حتى Ω11. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ UNIVERSAL SOVEREIGN FABRICS ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Identity Fabric Authority Fabric Knowledge Fabric Contract Fabric Execution Fabric Evidence Fabric Provenance Fabric Security Fabric Trust Fabric Determinism Fabric Evolution Fabric Communication Fabric Observation Fabric Verification Fabric ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ هذه الأقمشة (Fabrics) تمر رأسياً عبر جميع الطبقات، فلا تصبح الحوكمة أو الأمن أو الأدلة محصورة في حضارة واحدة. ⸻ Ω Sovereign Control Plane بين الدستور والتنفيذ ينقصك مستوى تحكم عالمي. Ω Constitution │ Sovereign Control Plane │ Policy Engine Authority Engine Capability Manager Resource Governor Admission Controller Constitutional Court Evolution Controller هذا المستوى هو العقل الإداري الذي ينسق المنظومة بأكملها. ⸻ Ω Canonical Registry لديك عدة Registries موزعة، والأفضل توحيدها تحت مرجع واحد. Canonical Registry ├── Identity Registry ├── Contract Registry ├── Module Registry ├── Runtime Registry ├── Knowledge Registry ├── Evidence Registry ├── Policy Registry ├── Capability Registry ├── Version Registry ├── Evolution Registry └── Artifact Registry فيصبح كل شيء يُكتشف من مكان واحد. ⸻ Ω Sovereign Bus بدلاً من ربط الطبقات مباشرة: VV ───────► VX V ───────► VX XV ─────► VX الأفضل: Sovereign Event Bus Events Commands Queries Streams Notifications Synchronization جميع الحضارات تتخاطب عبره. ⸻ Ω Digital DNA كل Artifact داخل النظام يمتلك DNA ثابتًا. Artifact DNA Identity Version Authority Lineage Evidence Dependencies Capabilities Security Profile Knowledge Profile Execution Profile Evolution History وبذلك يصبح أي عنصر قابلاً للتتبع والتحقق دون استثناء. ⸻ Ω Verification Mesh بدلاً من أن تكون طبقة تحقق واحدة: Runtime │ Verification │ Review OS اجعل التحقق شبكة موزعة: Identity Validator Contract Validator Knowledge Validator Runtime Validator Security Validator Policy Validator Ledger Validator Evidence Validator Evolution Validator Infrastructure Validator كل عقدة تتحقق من مجالها باستمرار. ⸻ Ω Civilization Lifecycle لكل حضارة دورة حياة موحدة: Genesis Specification Implementation Verification Deployment Operation Observation Evolution Deprecation Archive فتصبح جميع الحضارات تعمل بنفس النمط. ⸻ Ω Universal Interfaces أضف واجهات معيارية بين الحضارات: Sovereign Identity Interface Sovereign Contract Interface Knowledge Interface Runtime Interface Ledger Interface Review Interface Evolution Interface Security Interface وهذا يقلل الاقتران ويجعل كل حضارة قابلة للاستبدال دون كسر البقية. ⸻ Ω Topology بدلاً من رسم خطي: Constitution │ Sovereign Control Plane │ ══════════════════════════════════════ Universal Cross-Layer Fabrics ══════════════════════════════════════ V VV VPI \ │ / Sovereign Bus │ VX │ XV │ Review & Verification Mesh │ Specialized Civilizations Ω0..Ω11 │ Evolution Registry │ Canonical Registry │ Knowledge Feedback │ Constitution هذا يجعل الرسم دائريًا مغلقًا بدلاً من كونه تسلسلاً نزوليًا فقط. ⸻ النتيجة بعد هذه الإضافات، لا يعود الرسم مجرد مخطط طبقات أو هيكل مستودع، بل يصبح مرجعًا معماريًا يصف: * السلطة (Authority): من يقرر؟ * المعرفة (Knowledge): ماذا نعرف؟ * التنفيذ (Execution): كيف يعمل؟ * التحقق (Verification): كيف نثبت صحته؟ * التطور (Evolution): كيف يتغير؟ * الترابط (Connectivity): كيف تتواصل الحضارات؟ * التتبع (Traceability): كيف نتتبع كل شيء؟ * الهوية (Identity): كيف نعرّف كل كيان؟ * الحوكمة (Governance): كيف تُفرض القواعد؟ بهذا يتحول الرسم إلى Master Architecture Blueprint متعدد الأبعاد، حيث تُعرض الحضارات أفقيًا، بينما تمر الأقمشة السيادية والواجهات المشتركة رأسيًا عبر جميعها، مع حلقة تغذية راجعة مغلقة تحفظ الاتساق والتطور طويل الأمد.══════════════════════════════════════════════════════════╗ ║ VAIXLNS — CANONICAL META CIVILIZATION (MASTER RUNTIME) ║ ║ Sovereign Self-Evolving Civilization Computing Platform ║ ╚══════════════════════════════════════════════════════════════════════════════════════════════╝ Ω∞ META CIVILIZATION │ ▼ Ω0 CONSTITUTION CORE Immutable Laws • Identity • Invariants │ ▼ META THEORY ENGINE Logic • Type Theory • Category Theory • Proof Theory │ ▼ CANONICAL MEANING ENGINE (CME) Concepts • Semantics • Truth • Relations • Derivations • Versions │ ▼ CANONICAL TRUTH MODEL (CTM) Truth ≠ State • Formal Reality • Truth Graph │ ▼ CANONICAL DERIVATION GRAPH (CDG) Every Concept / Decision / Law / Component has Proven Origin │ ▼ ══════════════════════════════════════════════════════════════════════════════════════════════ SOVEREIGN CIVILIZATION CORE ══════════════════════════════════════════════════════════════════════════════════════════════ │ ┌──────────────────────────────┼──────────────────────────────┐ ▼ ▼ ▼ Knowledge Civilization Authority Civilization Execution Civilization │ │ │ ▼ ▼ ▼ Knowledge Fabric Governance Fabric VX Runtime Kernel Evidence Fabric Policy Engine Runtime Engine Meaning Graph Authority Algebra Event Runtime Concept Registry Trust Runtime State Runtime Truth Graph Constraint Calculus Scheduler Canonical Memory Constitutional Guard Execution Engine │ │ │ └──────────────────────────────┼──────────────────────────────┘ ▼ ══════════════════════════════════════════════════════════════════════════════════════════════ VERIFICATION CIVILIZATION ══════════════════════════════════════════════════════════════════════════════════════════════ │ ▼ Proof-Carrying Execution │ ┌────────────────────────────────┼────────────────────────────────┐ ▼ ▼ ▼ Assurance Fabric Evidence Verification Determinism Engine Identity Validator Proof Validator Replay Engine Contract Validator Reality Validator State Validation Capability Validator Governance Validator Integrity Proof Security Validator Runtime Validator Merkle Verification │ ▼ Sovereign Proof Package │ ▼ ══════════════════════════════════════════════════════════════════════════════════════════════ EVOLUTION CIVILIZATION ══════════════════════════════════════════════════════════════════════════════════════════════ │ ▼ Canonical Evolution Calculus │ ▼ Evolution Engine (SER) │ ┌─────────────────────────────────┼─────────────────────────────────┐ ▼ ▼ ▼ Gap Analyzer Strategy Planner Architecture Synthesizer │ │ │ └─────────────────────────────────┼─────────────────────────────────┘ ▼ ASGF Civilization │ ┌───────────────────────────────────┼───────────────────────────────────┐ ▼ ▼ ▼ Specification Engine Code Generator Documentation Generator Architecture Generator Test Generator Validation Generator Contract Generator Deployment Generator Runtime Generator │ ▼ ══════════════════════════════════════════════════════════════════════════════════════════════ ADAPTIVE EXECUTION FABRIC (AEF) ══════════════════════════════════════════════════════════════════════════════════════════════ │ Runtime Strategy Selection │ Load Analysis ─ Risk Analysis ─ Policy Selection ─ Optimization │ ▼ Deterministic Execution │ ▼ ══════════════════════════════════════════════════════════════════════════════════════════════ DEPLOYMENT FABRIC ══════════════════════════════════════════════════════════════════════════════════════════════ │ ▼ GROT ROOT OF TRUTH │ Registry • Version • Proof • Lineage • Capability │ ▼ Living Civilization Memory │ ▼ Continuous Evolution Loop │ └───────────────┐ │ ▼ META CIVILIZATION Ω∞رجعت للفهرس الكامل السابق. المشكلة ليست في فكرة VAIXLNS، المشكلة أن الحزم الأخيرة دخلت فوق بعضها بدون مرجع مركزي. قبل أي توليد جديد لازم نثبت Genesis Index + Build Pipeline ونربط كل الابتكارات السابقة داخله. سأعيد ترتيبها كحزمة واحدة (نسخة دمج)، تشمل: VAIXLNS GENESIS MASTER INDEX v0.3 0. المبدأ الأساسي V = Governance VX = Deterministic Execution XV = Intelligence / Prediction VAMM = Meta Model + Compiler SRE Ω = Stability + Recovery + Evolution LNS = Knowledge + Evidence Memory AURX = Network Fabric القانون: NO EVIDENCE ↓ NO GOVERNANCE FINALITY ↓ NO EXECUTION ⸻ 1. Core Runtime src/core/ ├── kernel.ts ├── atomic-executor.ts ├── ledger.ts ├── rollback-manager.ts ├── state-machine.ts ├── event-bus.ts └── deterministic-runtime.ts المسؤولية: * Event يدخل * Decision ينتج * Ledger يسجل * State يتغير * Rollback متوفر ⸻ 2. VX Kernel VX Kernel العقد: Event ↓ Validation ↓ Governance Check ↓ Finality Certificate ↓ Atomic Execution ↓ Ledger Commit ممنوع: Direct Mutation Randomness External Side Effect إلا عبر Ledger. ⸻ 3. Ledger System تعديل الجدول: CREATE TABLE ledger ( id INTEGER PRIMARY KEY, event_id TEXT, type TEXT, payload TEXT, prev_hash TEXT, hash TEXT, timestamp TEXT ); السلسلة: BLOCK1 hash=A BLOCK2 prev_hash=A hash=B BLOCK3 prev_hash=B hash=C ⸻ 4. VAMM Compiler المسار: YAML Artifact ↓ Parser ↓ Semantic Analyzer ↓ IR ↓ Planner ↓ Generator ↓ Verifier ↓ Runtime المجلد: vamm/ ├── meta-model.yaml ├── master-index.yaml ├── artifacts/ ├── blueprints/ └── templates/ ⸻ 5. VAMM Artifact Model كل شيء يصبح Artifact: id: kind: version: contract: implementation: tests: lifecycle: evidence: الأنواع: Specification Capability Runtime Agent Policy Decision Evidence Event Test Release ⸻ 6. Agent Factory المولد: Artifact | | Blueprint | | Agent Contract | | Generated Agent | | Evidence Hash الناتج: generated/ agents/ VV-Agent.ts AA-Agent.ts II-Agent.ts XX-Agent.ts *.evidence.yaml ⸻ 7. SRE Ω Layer Ω.70.08 Stability Engine المكونات: Volatility Drift Directionality Consistency الناتج: Stability Score يذهب إلى: Sphere PSE Collapse Detection Governance ⸻ 8. Predictive Engine Ω.70.10 وظيفته: Current State ↓ Temporal Analysis ↓ Future State Probability المخرجات: prediction: risk: confidence: trajectory: ⸻ 9. Collapse Detection Ω.70.09 يراقب: High Volatility High Drift Negative Direction Low Consistency Low Stability Score الناتج: Collapse Event ⸻ 10. Recovery Engine Ω.70.12 بعد الانهيار: Incident ↓ Diagnosis ↓ Recovery Plan ↓ Evidence Token ⸻ 11. Self Healing Manifest Ω.70.13 المسار: Incident ↓ Manifest Engine ↓ Improvement Proposal ↓ Simulation ↓ Governance ↓ Compiler ↓ Atomic Patch ⸻ 12. Governance Finality Finality Gate: if( entropy > threshold ) reject(); if( evidence missing ) reject(); if( quorum failed ) reject(); FINALIZED الناتج: FinalityCertificate ⸻ 13. Entropy Bound Finality إضافة جديدة: النظام لا يغير نفسه أثناء الفوضى. قاعدة: High Entropy = No Evolution ⸻ 14. Atomic Evolution Loop الدورة الكاملة: Observe ↓ Analyze ↓ Predict ↓ Recover ↓ Generate Improvement ↓ Simulate ↓ Govern ↓ Compile ↓ Atomic Swap ↓ Verify ↓ Learn ⸻ 15. Repository النهائي VAIXLNS/ ├── vamm/ │ │── meta-model.yaml │── master-index.yaml │ ├── src/ │ │── core/ │ ├── kernel.ts │ ├── ledger.ts │ ├── atomic-executor.ts │ │── vamm/ │ ├── compiler.ts │ ├── generator.ts │ │── agents/ │ │── sre/ │ ├── stability/ │ ├── predictive/ │ ├── collapse/ │ ├── recovery/ │ └── self-healing/ │ │── governance/ │ ├── finality.ts │ │── vx/ │ └── runtime.ts │ ├── generated/ │ ├── tests/ │ └── package.json ⸻ 16. التشغيل النهائي npm install npm run build npm run test npm run generate npm run genesis النتيجة: ✔ VAMM Loaded ✔ Semantic Graph Built ✔ Agents Generated ✔ Evidence Created ✔ Stability Online ✔ Governance Ready ✔ VX Runtime Started ✔ Genesis Loop Complete ⸻ هذا هو الفهرس المرجعي. أي ملف جديد يدخل بعد الآن يجب أن يرتبط بهذا المسار، وليس يضاف كفكرة منفصلة. الخطوة الصحيحة بعدها: توليد الحزمة التنفيذية الكاملة v0.3 ملفاً ملفاً بناءً على هذا الفهرس، وليس إضافة طبقات جديدة.أفهم طلبك: أنت لا تريد إعادة صياغة المعمارية، تريد فهرس تنفيذ (Execution Index) مضبوط يكون هو المرجع الذي تُبنى عليه الملفات والكود. الخطأ السابق كان أنه أعاد وصف الطبقات ولم يحوّلها إلى خطة بناء قابلة للتحويل إلى مستودع. هذا هو الفهرس التنفيذي الصحيح لـ VAIXLNS Genesis v0.3: VAIXLNS-GENESIS/ 00-CONTRACTS/ │ ├── system-contract.yaml ├── artifact-contract.yaml ├── event-contract.yaml ├── decision-contract.yaml ├── evidence-contract.yaml └── finality-contract.yaml 01-VAMM-COMPILER/ │ ├── meta-model.yaml ├── master-index.yaml │ ├── src/ │ ├── parser/ │ │ └── yaml-parser.ts │ │ │ ├── semantic/ │ │ └── semantic-analyzer.ts │ │ │ ├── ir/ │ │ └── artifact-ir.ts │ │ │ ├── planner/ │ │ └── generation-planner.ts │ │ │ ├── generator/ │ │ ├── generator.ts │ │ └── agent-factory.ts │ │ │ └── verifier/ │ └── compiler-verifier.ts 02-VX-RUNTIME/ │ ├── kernel/ │ ├── vx-kernel.ts │ ├── state-machine.ts │ └── execution-boundary.ts │ ├── events/ │ ├── event-bus.ts │ └── event-types.ts │ ├── decision/ │ └── decision-engine.ts │ └── runtime.ts 03-LEDGER/ │ ├── ledger.ts ├── hash-chain.ts ├── snapshot.ts ├── replay-engine.ts └── integrity-checker.ts 04-GOVERNANCE/ │ ├── finality-engine.ts ├── quorum-validator.ts ├── entropy-gate.ts ├── policy-engine.ts └── certificates/ 05-SRE-OMEGA/ │ ├── stability/ │ └── stability-engine.ts │ ├── predictive/ │ └── predictive-engine.ts │ ├── collapse/ │ └── collapse-detector.ts │ ├── recovery/ │ └── recovery-engine.ts │ └── self-healing/ ├── manifest-engine.ts ├── simulator.ts └── improvement-proposal.ts 06-AGENT-SYSTEM/ │ ├── agent-contract.ts ├── agent-runtime.ts ├── memory/ │ ├── ephemeral.ts │ ├── persistent.ts │ └── ledger-memory.ts │ └── generated/ 07-LNS-KNOWLEDGE/ │ ├── knowledge-store.ts ├── evidence-store.ts └── provenance-chain.ts 08-AURX-NETWORK/ │ ├── node-registry.ts ├── communication.ts └── discovery.ts 09-VERIFICATION/ │ ├── validator.ts ├── conformance-tests.ts ├── replay-tests.ts └── determinism-tests.ts 10-CLI/ │ ├── commands/ │ ├── build.ts │ ├── validate.ts │ ├── trace.ts │ ├── generate.ts │ └── genesis.ts │ └── cli.ts 11-TESTS/ │ ├── ledger.test.ts ├── kernel.test.ts ├── governance.test.ts ├── stability.test.ts ├── compiler.test.ts ├── agent-factory.test.ts └── end-to-end.genesis.test.ts 12-OUTPUT/ │ ├── generated-agents/ ├── generated-runtime/ ├── evidence/ ├── certificates/ └── reports/ ترتيب التنفيذ الإجباري: Phase 1 00-CONTRACTS 03-LEDGER 02-VX-RUNTIME Phase 2 04-GOVERNANCE 09-VERIFICATION Phase 3 05-SRE-OMEGA Phase 4 01-VAMM-COMPILER Phase 5 06-AGENT-SYSTEM Phase 6 10-CLI + Genesis Loop أول نسخة قابلة للتشغيل لا تبدأ بالـ Agent Factory ولا الـ Dashboard. تبدأ بهذه الثلاثة فقط: Ledger + VX Kernel + Governance Finality لأنها هي النواة التي تثبت أن النظام يعمل. بعد تثبيت هذا الفهرس، التنفيذ يكون ملفاً ملفاً بنفس المسارات، وليس توسعة أفكار جديدة.أنت محق في نقطة مهمة: الفهرس السابق ركّز على النواة التنفيذية فقط، لكنه أسقط الطبقات التي طلبتها سابقًا: واجهات التحكم، الوكلاء، اللغات، والمصنع الكامل. هذا ليس فهرس VAIXLNS الكامل. هذا هو الفهرس المصحح الذي يجمع Core + Control Plane + Agents + Languages + Compiler + UI: VAIXLNS-GENESIS/ 00-CONSTITUTION/ │ ├── genesis-manifest.yaml ├── sovereignty-rules.yaml ├── system-contracts/ │ ├── artifact-contract.yaml │ ├── agent-contract.yaml │ ├── event-contract.yaml │ ├── decision-contract.yaml │ └── evidence-contract.yaml 01-VAMM-COMPILER/ │ ├── meta-model/ │ └── meta-model.yaml │ ├── compiler/ │ ├── parser/ │ ├── semantic-engine/ │ ├── intermediate-representation/ │ ├── planner/ │ ├── generator/ │ └── verifier/ │ ├── languages/ │ ├── VQL/ │ │ ├── grammar.vql │ │ ├── parser.ts │ │ └── executor.ts │ │ │ ├── VIX/ │ │ ├── specification.vix │ │ ├── compiler.ts │ │ └── runtime.ts │ │ │ └── VASM/ │ ├── syntax │ └── assembler.ts 02-VX-SOVEREIGN-RUNTIME/ │ ├── kernel/ │ ├── vx-kernel.ts │ ├── scheduler.ts │ ├── decision-engine.ts │ └── execution-boundary.ts │ ├── state/ │ ├── state-machine.ts │ └── replay-engine.ts │ ├── event-system/ │ ├── event-bus.ts │ └── event-stream.ts │ └── runtime-manager.ts 03-AGENT-ECOSYSTEM/ │ ├── agent-runtime/ │ ├── agent-factory/ │ ├── agent-registry/ │ ├── agents/ │ │ ├── VV/ │ │ └── Verification-Vision-Agent │ │ │ ├── AA/ │ │ └── Analysis-Architecture-Agent │ │ │ ├── II/ │ │ └── Identity-Intelligence-Agent │ │ │ ├── XX/ │ │ └── Execution-Agent │ │ │ ├── LL/ │ │ └── Ledger-Agent │ │ │ ├── NN/ │ │ └── Network-Agent │ │ │ └── SS/ │ └── Security-Agent │ └── agent-memory/ 04-SRE-OMEGA/ │ ├── stability-engine/ ├── predictive-engine/ ├── collapse-detection/ ├── recovery-engine/ ├── self-healing/ └── evolution-engine/ 05-GOVERNANCE-CONTROL-PLANE/ │ ├── governance-engine/ ├── finality-engine/ ├── policy-engine/ ├── certificate-manager/ │ ├── dashboards/ │ │ ├── Sovereign Dashboard │ │ │ ├── Runtime Dashboard │ │ │ ├── Stability Dashboard │ │ │ ├── Agent Fleet Dashboard │ │ │ ├── Evidence Dashboard │ │ │ └── Governance Approval Dashboard │ └── human-console/ 06-FLIGHT-DECK/ │ ├── web-ui/ │ ├── mobile-console/ │ ├── realtime-monitor/ │ ├── graph-viewer/ │ ├── timeline-viewer/ │ └── incident-console/ 07-LNS-KNOWLEDGE-SYSTEM/ │ ├── knowledge-engine/ ├── semantic-memory/ ├── evidence-store/ ├── provenance-chain/ └── institutional-memory/ 08-AURX-NETWORK/ │ ├── node-fabric/ ├── service-discovery/ ├── secure-channel/ └── network-orchestration/ 09-DATA-PLATFORM/ │ ├── ledger/ ├── database/ ├── vector-store/ ├── event-store/ └── telemetry/ 10-OBSERVABILITY/ │ ├── OpenTelemetry ├── Metrics ├── Logs ├── Traces ├── Causal-Debugger └── Reality-Map 11-SECURITY/ │ ├── identity/ ├── encryption/ ├── secrets/ ├── threat-detection/ └── shield-engine/ 12-DEV-FACTORY/ │ ├── repository-generator ├── code-generator ├── test-generator ├── documentation-generator └── release-generator 13-CLI/ │ ├── vamm ├── vx ├── sre ├── agent ├── governance └── genesis 14-TESTING/ │ ├── unit-tests ├── integration-tests ├── simulation-tests ├── determinism-tests ├── chaos-tests └── proof-tests الآن الصورة كاملة: * الشاشات موجودة → Flight Deck + Control Plane * الوكلاء موجودون → VV / AA / II / XX / LL / NN / SS * اللغات موجودة → VQL + VIX + VASM * المترجم موجود → VAMM Compiler * النواة موجودة → VX Runtime * الذاكرة موجودة → LNS * التشبيك موجود → AURX * الاستقرار والتطور موجود → SRE Ω * التوليد الآلي موجود → Dev Factory لكن التنفيذ الصحيح لا يبدأ بكتابة كل هذا دفعة واحدة؛ يبدأ ببناء Genesis v0.1 قابل للتشغيل ثم تتولد بقية الطبقات من VAMM. أي محاولة لكتابة كل المجلدات مرة واحدة ستنتج هيكل ضخم غير قابل للاختبار.# GitHub's plans An overview of GitHub's pricing plans. ## About GitHub's plans GitHub offers free and paid plans for storing and collaborating on code. Some plans are available only to personal accounts, while other plans are available only to organization and enterprise accounts. For more information about accounts, see Types of GitHub accounts. See costs and features for each plan at GitHub Pricing. For information on planned features and products, see the GitHub public roadmap. When you read GitHub Docs, make sure to select the version that reflects your plan. For more information, see About versions of GitHub Docs. ## GitHub Free for personal accounts With GitHub Free for personal accounts, you can work with unlimited collaborators on unlimited public repositories with a full feature set, and on unlimited private repositories with a limited feature set. With GitHub Free, your personal account includes: * GitHub Community Support * Dependabot alerts * Deployment protection rules for public repositories * Two-factor authentication enforcement * 500 MB GitHub Packages storage * GitHub Actions features: * 2,000 minutes per month * Deployment protection rules for public repositories * GitHub Pages in public repositories ## GitHub Pro In addition to the features available with GitHub Free for personal accounts, GitHub Pro includes: * GitHub Support via email * 3,000 GitHub Actions minutes per month * 2 GB GitHub Packages storage * Advanced tools and insights in private repositories: * Required pull request reviewers * Multiple pull request reviewers * Protected branches * Code owners * Auto-linked references * GitHub Pages > [!NOTE] > To publish a GitHub Pages site privately, you need to have an organization account. Additionally, your organization must use GitHub Enterprise Cloud. * Wikis * Repository insights graphs: Pulse, contributors, traffic, commits, code frequency, network, and forks ## GitHub Free for organizations With GitHub Free for organizations, you can work with unlimited collaborators on unlimited public repositories with a full feature set, or unlimited private repositories with a limited feature set. In addition to the features available with GitHub Free for personal accounts, GitHub Free for organizations includes: * GitHub Community Support * Team access controls for managing groups * 2,000 GitHub Actions minutes per month ## GitHub Team In addition to the features available with GitHub Free for organizations, GitHub Team includes: * GitHub Support via email * 3,000 GitHub Actions minutes per month * 2 GB GitHub Packages storage * The option to purchase GitHub Advanced Security products: * GitHub Code Security * GitHub Secret Protection For more information, see About GitHub Advanced Security. * Advanced tools and insights in private repositories: * Required pull request reviewers * Multiple pull request reviewers * Team pull request reviewers * Protected branches * Code owners * Scheduled reminders * GitHub Pages > [!NOTE] > To publish a GitHub Pages site privately, you need to have an organization account. Additionally, your organization must use GitHub Enterprise Cloud. * Wikis * Security overview * Repository insights graphs: Pulse, contributors, traffic, commits, code frequency, network, and forks GitHub bills for GitHub Team on a per-user basis. For more information, see About per-user pricing in the Free, Pro, & Team documentation. GitHub Actions usage is free for standard GitHub-hosted runners in public repositories, and for self-hosted runners. See Choosing the runner for a job. For private repositories, each GitHub account receives a quota of free minutes and storage for use with GitHub-hosted runners, depending on the account's plan. Any usage beyond the included amounts is billed to your account. ## GitHub Enterprise GitHub Enterprise includes two deployment options: GitHub Enterprise Cloud, which is hosted by GitHub in the cloud, and GitHub Enterprise Server, which is self-hosted. For more information, see About GitHub for enterprises. In addition to the features available with GitHub Team, GitHub Enterprise includes: * GitHub Enterprise Support * Additional security, compliance, and deployment controls * Authentication with SAML single sign-on * Access provisioning with SAML or SCIM * Deployment protection rules with GitHub Actions for private or internal repositories * GitHub Connect * Additional features such as internal repositories and repository rules. GitHub Enterprise Cloud specifically includes: * 50,000 GitHub Actions minutes per month * Included minutes can be used with standard GitHub-hosted runners only. For more information about billing for larger runners, see Larger runners. * 50 GB GitHub Packages storage * A service level agreement for 99.9% monthly uptime * The option to centrally manage policy and billing for multiple GitHub organizations with an enterprise account. For more information, see Enterprise accounts. * If you upgrade an existing organization to GitHub Enterprise Cloud, your enterprise account will contain your organization. For more information, see Upgrading your account's plan and Creating an enterprise account in the GitHub Enterprise Cloud documentation. * The option to provision and manage the user accounts for your developers, by using Enterprise Managed Users. For more information, see About Enterprise Managed Users. * Additional features such as audit log streaming and IP allow list. For more information, see Streaming the audit log for your enterprise and Restricting network traffic to your enterprise with an IP allow list in the GitHub Enterprise Cloud documentation. * The option to host your company's data in a specific region, on a unique subdomain For a detailed list of the features available with GitHub Enterprise, see our Pricing page. To see how GitHub compares to GitLab, Bitbucket, and Jenkins, see Comparing GitHub to other DevOps solutions. You can set up trials to evaluate GitHub Enterprise Cloud and GitHub Enterprise Server. For more information, see Setting up a trial of GitHub Enterprise Cloud and Setting up a trial of GitHub Enterprise Server. ## Further reading * About per-user pricing in the GitHub Enterprise Cloud documentationأقدر الهدف، لكن لا أستطيع أن أقول “تم التنفيذ الكامل” أو “كل البراهين اكتملت” لأن ذلك سيكون ادعاءً غير صحيح. ما تملكه الآن هو مواصفة معمارية (Specification) متقدمة جداً، وليست نواة مبرهنة أو منفذة بعد. إذا أردنا تنفيذًا كاملاً، فسيكون على مراحل واضحة، لأن بناء نظام بهذا الحجم يعني كتابة عشرات الآلاف من الأسطر، ومئات المبرهنات في Lean 4، وآلاف اختبارات المطابقة. التنفيذ الحقيقي سيكون بهذا التسلسل: Ω Specification │ ▼ Specification Dependency Graph (SDG) │ ▼ Civilization Compiler │ ├──────────────┐ │ │ ▼ ▼ Formal Generator Runtime Generator (Lean 4) (Rust) │ │ ▼ ▼ Proof Objects Runtime Modules │ │ └──────┬───────┘ ▼ ABI Generator │ ▼ Conformance Generator │ ▼ Executable Runtime │ ▼ Self Validation Loop │ ▼ Evidence → Lineage → Registry ما الذي سيُولد فعلياً؟ ليس ملفًا واحدًا، بل مشروعًا كاملاً: vaixlns/ meta/ constitution/ invariants/ registry/ ontology/ sdg/ compiler/ parser/ semantic/ validator/ planner/ dependency/ generators/ rust/ lean/ abi/ docs/ tests/ schemas/ sdk/ runtime/ kernel/ scheduler/ authority/ capability/ trust/ resource/ runtime_physics/ networking/ ledger/ projection/ recovery/ proofs/ state/ preservation/ progress/ authority/ effects/ scheduler/ determinism/ replay/ conformance/ replay/ snapshot/ rollback/ property/ abi/ serialization/ evidence/ proof/ replay/ merkle/ authority/ lineage/ factories/ kernel_factory/ runtime_factory/ service_factory/ sdk_factory/ clone_factory/ civilization_factory/ evolution/ simulator/ benchmark/ strategy/ optimizer/ bootstrap/ genesis/ bootloader/ root/ ⸻ ثم يبدأ التنفيذ الذاتي Meta Registry ↓ Compiler ↓ Dependency Graph ↓ Planner ↓ Generator ↓ Rust ↓ Lean ↓ ABI ↓ Tests ↓ Runtime ↓ Evidence ↓ Proof ↓ Certification ↓ Evolution ↓ Meta Registry ⸻ عندها يصبح النظام قادراً على توليد: * Kernel * Runtime * Services * SDK * APIs * Documentation * Tests * Schemas * Proofs * Evidence * ABI * WASM * Bindings * Registries * Lineage * Governance * Deployments كلها من مصدر واحد هو الـ Canonical Meta Model. لكن هناك نقطة مهمة حتى مع اكتمال هذه المعمارية، لن تكون البراهين موجودة تلقائياً. لا بد من: 1. كتابة الـ SDG الحقيقي. 2. كتابة الـ Canonical Meta Model. 3. بناء Civilization Compiler. 4. توليد كود Rust. 5. توليد ملفات Lean 4. 6. إكمال وإثبات المبرهنات داخل Lean. 7. تشغيل اختبارات المطابقة. 8. إصدار الأدلة (Evidence). 9. الحصول على حالة Verified/Certified. هذا هو التنفيذ الكامل الواقعي. أي ادعاء بأن كل ذلك قد اكتمل الآن سيكون غير دقيق. حالياً أنتم في مرحلة تصميم مواصفة غنية، والخطوة التالية هي تحويلها إلى مشروع برمجي ومنظومة برهانية فعلية.أتفق مع الفكرة الأساسية، لكن مع تصحيح مهم: ما وصلتم إليه ليس نهاية التصميم بالمعنى الهندسي، بل هو نهاية المعمارية المرجعية (Reference Architecture). بعد ذلك تبدأ دورة طويلة من التنفيذ، والتحقق، والقياس، وإعادة التصميم الجزئي. حتى الأنظمة الكبرى مثل Linux وPostgreSQL وRust نفسها استمرت في التطور بعد اكتمال معماريتها الأساسية. إذا أردنا تنفيذ VAIXLNS وفق الرسم الهندسي الحالي دون حذف أي مفهوم أو تغيير أي طبقة، فإن التحويل الطبيعي يكون بهذا الشكل: VAIXLNS Ω∞ → Repository Skeleton vaixlns/ │ ├─────────────────────────────────────────────── │ Ω∞ META FIELD ├─────────────────────────────────────────────── │ ├── omega/ │ ├── ontology/ │ ├── algebra/ │ ├── semantics/ │ ├── mathematics/ │ ├── logic/ │ ├── proof/ │ ├── specification/ │ ├── theorem/ │ └── invariants/ │ ├─────────────────────────────────────────────── │ Ω0 GENESIS CONSTITUTION ├─────────────────────────────────────────────── │ ├── genesis/ │ ├── vaixlns_root.lns │ ├── parser/ │ ├── validator/ │ ├── constitution/ │ ├── identity/ │ ├── authority/ │ ├── domains/ │ ├── entities/ │ ├── relations/ │ ├── laws/ │ └── evolution_rules/ │ ├─────────────────────────────────────────────── │ VX SOVEREIGN KERNEL ├─────────────────────────────────────────────── │ ├── kernel/ │ ├── identity_kernel/ │ ├── event_kernel/ │ ├── time_kernel/ │ ├── crypto_kernel/ │ ├── capability_kernel/ │ ├── resource_kernel/ │ ├── execution_context/ │ ├── runtime_context/ │ └── kernel_api/ │ ├─────────────────────────────────────────────── │ DCEG EXECUTION FABRIC ├─────────────────────────────────────────────── │ ├── dceg/ │ ├── resolver/ │ ├── dependency_graph/ │ ├── scheduler/ │ ├── logical_workers/ │ ├── conflict_analyzer/ │ ├── retry_policy/ │ ├── quarantine/ │ ├── deterministic_queue/ │ └── execution_engine/ │ ├─────────────────────────────────────────────── │ MVCC STATE FABRIC ├─────────────────────────────────────────────── │ ├── mvcc/ │ ├── snapshot/ │ ├── state/ │ ├── rw_sets/ │ ├── validator/ │ ├── version_store/ │ ├── transaction/ │ ├── serializer/ │ └── replay_snapshot/ │ ├─────────────────────────────────────────────── │ DIGITAL TWIN VALIDATION ├─────────────────────────────────────────────── │ ├── twin/ │ ├── simulator/ │ ├── shadow_runtime/ │ ├── stress/ │ ├── security/ │ ├── fuzzing/ │ ├── deterministic_replay/ │ ├── chaos/ │ └── validator/ │ ├─────────────────────────────────────────────── │ CONSTITUTIONAL COMMIT ├─────────────────────────────────────────────── │ ├── commit/ │ ├── invariant_engine/ │ ├── constitutional_engine/ │ ├── policy_engine/ │ ├── logic_engine/ │ ├── constraint_solver/ │ ├── commit_barrier/ │ ├── arbitration/ │ └── approval/ │ ├─────────────────────────────────────────────── │ PROOF FABRIC ├─────────────────────────────────────────────── │ ├── proof/ │ ├── vbin/ │ ├── hash/ │ ├── merkle/ │ ├── signatures/ │ ├── replay_hash/ │ ├── certificates/ │ ├── proof_chain/ │ └── verifier/ │ ├─────────────────────────────────────────────── │ EXL MEMORY FABRIC ├─────────────────────────────────────────────── │ ├── exl/ │ ├── ledger/ │ ├── evidence/ │ ├── storage/ │ ├── snapshots/ │ ├── replay/ │ ├── compression/ │ ├── indexing/ │ └── archive/ │ ├─────────────────────────────────────────────── │ EVOLUTION ENGINE ├─────────────────────────────────────────────── │ ├── evolution/ │ ├── observer/ │ ├── analyzer/ │ ├── planner/ │ ├── synthesizer/ │ ├── simulator/ │ ├── proof_gate/ │ ├── proposer/ │ └── constitutional_upgrade/ │ ├─────────────────────────────────────────────── │ DISTRIBUTED FABRIC ├─────────────────────────────────────────────── │ ├── distributed/ │ ├── cluster/ │ ├── node/ │ ├── replication/ │ ├── consensus/ │ ├── discovery/ │ ├── routing/ │ ├── transport/ │ └── synchronization/ │ ├─────────────────────────────────────────────── │ OBSERVABILITY ├─────────────────────────────────────────────── │ ├── observability/ │ ├── metrics/ │ ├── tracing/ │ ├── logging/ │ ├── alerts/ │ ├── profiling/ │ └── diagnostics/ │ ├─────────────────────────────────────────────── │ SELF CERTIFICATION ├─────────────────────────────────────────────── │ ├── certification/ │ ├── runtime_check/ │ ├── replay_check/ │ ├── invariant_check/ │ ├── proof_check/ │ ├── governance_check/ │ ├── certificate_builder/ │ └── release_certificate/ │ └── runtime/ └── main.rs ⸻ طبقة الـ Traits (العقود) لا تتخاطب أي طبقة مباشرة مع الأخرى، بل عبر واجهات ثابتة: Genesis │ ▼ ConstitutionProvider Kernel │ ▼ ExecutionProvider DCEG │ ▼ SchedulingProvider MVCC │ ▼ StateProvider Twin │ ▼ ValidationProvider Commit │ ▼ CommitProvider Proof │ ▼ ProofProvider EXL │ ▼ LedgerProvider Evolution │ ▼ EvolutionProvider Distributed │ ▼ ClusterProvider هذا يحافظ على العزل الكامل بين الطبقات ويجعل استبدال أي تنفيذ داخلي ممكناً دون كسر بقية النظام. ⸻ أول إثباتات رياضية (Formal Core) قبل كتابة آلاف الأسطر من Rust، تكون أول البديهيات الرسمية: Axiom 1 Genesis Immutable ∀ t Genesis(t)=Genesis(0) Axiom 2 Replay Replay(trace) = Original Execution Axiom 3 Commit Accepted ⇔ Constitution ∧ Proof ∧ No Conflict ∧ Replay Valid Axiom 4 EXL Hash(n) = H( Hash(n−1), Evidence(n) ) Axiom 5 Determinism Input₁ = Input₂ ↓ Output₁ = Output₂ ⸻ أول Pipeline تنفيذي Developer │ ▼ Genesis Parser │ ▼ Ontology Validator │ ▼ Kernel Boot │ ▼ DCEG Scheduler │ ▼ MVCC Snapshot │ ▼ Digital Twin │ ▼ Commit Barrier │ ▼ Proof Fabric │ ▼ EXL Ledger │ ▼ Replay Verification │ ▼ Self Certification │ ▼ Distributed Commit التقييم الهندسي إذا نُفِّذت هذه الطبقات كما هي، مع تنفيذ فعلي وليس مجرد هياكل فارغة، فسيكون الناتج أقرب إلى منصة تشغيل موزعة ذات خصائص تحقق قوية وليس مجرد “إطار عمل ذكاء اصطناعي”. لكن الوصول إلى مستوى إنتاجي يتطلب سنوات من العمل على التنفيذ، والتحقق الرسمي، والاختبارات، وتحسين الأداء، وإدارة التعقيد. المعمارية وحدها لا تضمن ذلك، لكنها توفر أساسًا منظمًا يمكن البناء عليه.VX – Canonical VAIXLNS_ROOT.lns + Single‑File Genesis Engine (v1.0) نولّعها من الجذر، بدون التفاف. أولاً: الدستور الحضاري (VAIXLNS_ROOT.lns). ثانياً: النواة التنفيذية السيادية (Rust single‑file Genesis). --- 1. Canonical VAIXLNS_ROOT.lns vaixlns_root { meta { id "VAIXLNS://ROOT/GENESIS/1.0" name "VAIXLNS Civilization Root" version "1.0.0" authority "VX_SOVEREIGN_CORE" } ontology { domains { core "Civilization Core" memory "Living Memory System" graph "Dependency & Knowledge Graph" constitution "Sovereign Constitutional Layer" rce "Runtime Control & Execution" economy "Resource & Value Economy" agents "Autonomous Sovereign Agents" orchestrator "Global Orchestration Layer" runtime "Execution Substrate" } entities { core.kernel { domain core kind "kernel" provides ["identity", "timebase", "event_bus"] requires ["constitution.root", "graph.root"] } memory.store { domain memory kind "living_store" provides ["state", "history", "snapshots"] requires ["core.kernel"] } graph.root { domain graph kind "graph_root" provides ["dependency_graph", "knowledge_graph"] requires ["core.kernel"] } constitution.root { domain constitution kind "sovereign_constitution" provides ["rules", "constraints", "sovereign_signatures"] requires ["core.kernel", "graph.root"] } rce.controller { domain rce kind "runtime_controller" provides ["execution_control", "scheduling", "supervision"] requires ["core.kernel", "constitution.root"] } economy.engine { domain economy kind "value_engine" provides ["credits", "quotas", "resource_policies"] requires ["core.kernel", "memory.store"] } agents.mesh { domain agents kind "agent_mesh" provides ["sovereign_agents", "roles", "capabilities"] requires ["core.kernel", "memory.store", "graph.root"] } orchestrator.global { domain orchestrator kind "global_orchestrator" provides ["flows", "pipelines", "coordination"] requires ["core.kernel", "agents.mesh", "rce.controller"] } runtime.substrate { domain runtime kind "execution_substrate" provides ["threads", "tasks", "io"] requires ["rce.controller"] } } relations { link core.kernel -> memory.store link core.kernel -> graph.root link core.kernel -> constitution.root link core.kernel -> rce.controller link core.kernel -> economy.engine link core.kernel -> agents.mesh link core.kernel -> orchestrator.global link rce.controller -> runtime.substrate link orchestrator.global -> agents.mesh link orchestrator.global -> economy.engine } } laws { sovereignty { rule "no_unsigned_transition" { require "constitution.root.sovereign_signatures" enforce "all_state_changes_signed" } rule "no_orphan_entity" { enforce "every_entity_has_domain_and_dependencies" } } consistency { rule "acyclic_dependency_graph" { enforce "graph.root.dependency_graph.is_acyclic" } rule "no_broken_links" { enforce "all_relations.resolve_successfully" } } lineage { rule "immutable_genesis" { enforce "vaixlns_root.meta.version_is_genesis" } } } axes { axis "sovereignty" { weight 1.0 binds ["constitution.root", "core.kernel", "agents.mesh"] } axis "memory" { weight 0.9 binds ["memory.store", "graph.root"] } axis "execution" { weight 0.9 binds ["rce.controller", "runtime.substrate", "orchestrator.global"] } axis "economy" { weight 0.8 binds ["economy.engine", "agents.mesh"] } } } --- 2. Single‑File Sovereign Genesis Engine (Rust) // vx_genesis.rs — Single-file Sovereign Reference v1.0 // ---------- Core Specs ---------- pub struct RootSpec { pub path: String, pub raw: String, } impl RootSpec { pub fn load(path: &str) -> Self { // هنا لاحقاً: قراءة الملف فعلياً Self { path: path.to_string(), raw: String::new(), } } } // ---------- Ontology ---------- pub struct Ontology { pub domains: Vec, pub entities: Vec, pub relations: Vec, pub laws: Laws, pub axes: Vec, } impl Default for Ontology { fn default() -> Self { Self { domains: Vec::new(), entities: Vec::new(), relations: Vec::new(), laws: Laws::default(), axes: Vec::new(), } } } pub struct Domain { pub id: String, pub name: String, } pub struct Entity { pub id: String, pub domain: String, pub kind: String, pub provides: Vec, pub requires: Vec, } pub struct Relation { pub from: String, pub to: String, } pub struct Laws { pub sovereignty: Vec, pub consistency: Vec, pub lineage: Vec, } impl Default for Laws { fn default() -> Self { Self { sovereignty: Vec::new(), consistency: Vec::new(), lineage: Vec::new(), } } } pub struct LawRule { pub id: String, pub description: String, } pub struct Axis { pub id: String, pub weight: f32, pub binds: Vec, } impl Ontology { pub fn parse(&mut self, root: &RootSpec) { // تفسير VAIXLNS_ROOT.lns إلى domains/entities/relations/laws/axes let _ = &root.raw; } pub fn validate(&self) { // تحقق من صحة الأنطولوجيا } pub fn expand(&mut self) { // توسعة الأنطولوجيا (استنتاج، ربط، إلخ) } } // ---------- Dependency Graph ---------- pub struct DependencyGraph { pub nodes: Vec, pub edges: Vec<(String, String)>, } impl Default for DependencyGraph { fn default() -> Self { Self { nodes: Vec::new(), edges: Vec::new(), } } } impl DependencyGraph { pub fn scan(&mut self, ontology: &Ontology) { for e in &ontology.entities { self.nodes.push(e.id.clone()); for req in &e.requires { self.edges.push((e.id.clone(), req.clone())); } } } pub fn resolve(&mut self) { // حل الاعتماديات، ترتيب، كشف الدورات } pub fn build_graph(&mut self) { // بناء تمثيل داخلي نهائي } } // ---------- Engines ---------- pub struct Engines { pub core: Option, pub memory: Option, pub graph: Option, pub constitution: Option, pub rce: Option, pub economy: Option, pub agents: Option, pub orchestrator: Option, pub runtime: Option, } impl Default for Engines { fn default() -> Self { Self { core: None, memory: None, graph: None, constitution: None, rce: None, economy: None, agents: None, orchestrator: None, runtime: None, } } } pub struct CoreEngine; pub struct MemoryEngine; pub struct GraphEngine; pub struct ConstitutionEngine; pub struct RceEngine; pub struct EconomyEngine; pub struct AgentsEngine; pub struct OrchestratorEngine; pub struct RuntimeEngine; impl Engines { pub fn create_core(&mut self, _ontology: &Ontology, _deps: &DependencyGraph) { self.core = Some(CoreEngine); } pub fn create_memory(&mut self) { self.memory = Some(MemoryEngine); } pub fn create_graph(&mut self) { self.graph = Some(GraphEngine); } pub fn create_constitution(&mut self) { self.constitution = Some(ConstitutionEngine); } pub fn create_rce(&mut self) { self.rce = Some(RceEngine); } pub fn create_economy(&mut self) { self.economy = Some(EconomyEngine); } pub fn create_agents(&mut self) { self.agents = Some(AgentsEngine); } pub fn create_orchestrator(&mut self) { self.orchestrator = Some(OrchestratorEngine); } pub fn create_runtime(&mut self) { self.runtime = Some(RuntimeEngine); } pub fn link_all(&mut self) { // ربط المحركات ببعضها وفق الأنطولوجيا والـ DependencyGraph } } // ---------- Sovereignty / Consistency / Lineage ---------- pub struct Sovereignty; pub struct Consistency; pub struct Lineage; impl Sovereignty { pub fn check(_engines: &Engines) { // تحقق من السيادة: لا انتقال غير موقّع، لا كيان بلا دستور } } impl Consistency { pub fn enforce(_engines: &Engines) { // فرض الاتساق: لا دورات، لا روابط مكسورة } } impl Lineage { pub fn verify(_engines: &Engines) { // تحقق من نسب الحضارة: genesis ثابت، المسار واضح } } // ---------- Genesis Engine ---------- pub struct GenesisEngine { root: RootSpec, ontology: Ontology, deps: DependencyGraph, engines: Engines, } impl GenesisEngine { pub fn new(root_path: &str) -> Self { Self { root: RootSpec::load(root_path), ontology: Ontology::default(), deps: DependencyGraph::default(), engines: Engines::default(), } } pub fn ignite(&mut self) { self.ontology.parse(&self.root); self.ontology.validate(); self.ontology.expand(); self.deps.scan(&self.ontology); self.deps.resolve(); self.deps.build_graph(); self.engines.create_core(&self.ontology, &self.deps); self.engines.create_memory(); self.engines.create_graph(); self.engines.create_constitution(); self.engines.create_rce(); self.engines.create_economy(); self.engines.create_agents(); self.engines.create_orchestrator(); self.engines.create_runtime(); self.engines.link_all(); self.verify(); self.launch(); } fn verify(&self) { Sovereignty::check(&self.engines); Consistency::enforce(&self.engines); Lineage::verify(&self.engines); } fn launch(&self) { // تشغيل الـ Runtime فعلياً // self.engines.runtime.as_ref().unwrap().start(); } } // ---------- Entry Point ---------- fn main() { let mut genesis = GenesisEngine::new("VAIXLNS_ROOT.lns"); genesis.ignite(); } --- لو حاب نرفعها لمرحلة Merkle Roots + توقيع سيادي فعلي داخل نفس الملف، نقدر نضيف طبقة Sovereign Signatures كـ module مدمج في هذا الـ single‑file… وتتحول ignite() إلى أول انتقال موقّع في تاريخ VX. 🚀🔥وبعدين؟ هنا تبدأ المرحلة التي تفرق بين فكرة ومنصة. VAIXLNS │ ▼ المرحلة 0 (Foundation) │ ▼ المرحلة 1 (Core) │ ▼ المرحلة 2 (Compiler) │ ▼ المرحلة 3 (Generator) │ ▼ المرحلة 4 (Runtime) │ ▼ المرحلة 5 (Evolution) │ ▼ المرحلة 6 (Distributed) │ ▼ المرحلة 7 (Civilization) ثم نبدأ نبني كل مرحلة حتى تصبح قابلة للتشغيل. لكن هناك شيء أهم من كل هذا. ⸻ أقترح تغيير طريقة البناء بالكامل بدلاً من بناء المشروع يدويًا، نبني أولًا: VAIXLNS Builder أي أن أول برنامج داخل VAIXLNS يكون مهمته بناء VAIXLNS نفسه. يعني: Mission │ ▼ Builder │ ▼ ينشئ Repository Folders Files Tests Contracts CI Documentation Examples ثم لاحقًا: Builder │ ▼ يبني VX XV ASGF SER GROT VV ثم: Builder │ ▼ يبني أي مشروع آخر ⸻ وهنا تبدأ الفكرة التي أعتقد أنها أقوى ابتكار يمكن أن يميزكم بدل أن يكون لديكم: Source Code يصبح لديكم: Civilization Blueprint │ ▼ VAIXLNS Builder │ ▼ Working System أي أن الكود يصبح ناتجًا، وليس الأصل. ⸻ لو كنت مهندس المشروع، فلن أبدأ بكتابة VX أو Runtime أولًا. سأبدأ بكتابة: VAIXLNS Builder v1 لأن بعد نجاحه، كل طبقة لاحقة يمكن أن تُولَّد وتُحدَّث من خلاله بدل بنائها يدويًا. هذه هي النقطة التي أراها قادرة على نقل VAIXLNS من مستودع كود إلى منصة هندسية تُنتج أنظمة قابلة للتشغيل والتطور.

• Review status: Below qualification threshold by 7.27 points; included because it ranked in today's Top 5.
• Score: 47.73/100
• Why PlotPickle should care: Agent Runtime / Context / Provider.
• What can PlotPickle learn from this? Investigate how the project approaches agent, context, memory, orchestration, runtime; repository metadata alone does not prove implementation fit.
• PlotPickle target: Agent Runtime / Context / Provider
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: UNKNOWN (unknown)
• Activity: active within 12 days
• Review history: first Radar review
• Score evidence: activity:recent, fork:no, lane:ai-architecture, license:unknown

4. test-zeus-ai/testzeus-hercules — WATCH

Hercules is the world’s first open-source testing agent, enabling UI, API, Security, Accessibility, and Visual validations – all without code or maintenance. Automate testing effortlessly and let Hercules handle the heavy lifting! ⚡

• Review status: Below qualification threshold by 18.42 points; included because it ranked in today's Top 5.
• Score: 46.58/100
• Why PlotPickle should care: Validation / Local Platform / Documentation.
• What can PlotPickle learn from this? Investigate how the project approaches automation, browser, testing, visual; repository metadata alone does not prove implementation fit.
• PlotPickle target: Validation / Local Platform / Documentation
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: AGPL-3.0 (review-required)
• Activity: last activity about 40 days ago
• Review history: first Radar review
• Score evidence: activity:recent, fork:no, lane:platform-engineering, license:review-required

5. Orkas-AI/Orkas — WATCH

Orkas is an open-source, local-first AI desktop app: a commander LLM directs specialist sub-agents, and runs your installed coding CLIs — Claude Code, Codex, OpenCode, OpenClaw, Hermes — as local sessions. Agents self-evolve via reflection and skill crystallization. BYO keys. macOS / Windows / Linux.

• Review status: Below qualification threshold by 19.48 points; included because it ranked in today's Top 5.
• Score: 45.52/100
• Why PlotPickle should care: Validation / Local Platform / Documentation.
• What can PlotPickle learn from this? Investigate how the project approaches app, desktop, local, windows; repository metadata alone does not prove implementation fit.
• PlotPickle target: Validation / Local Platform / Documentation
• Craft Module fit: none evidenced
• Related Issue fit: DEMO: Onboarding DEMO for New Users #1692
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: MIT (known-open-source)
• Activity: active within 1 days
• Review history: first Radar review
• Score evidence: activity:recent, fork:no, lane:platform-engineering, license:known-open-source

[github-actions]

OSS Radar — 2026-09-14

GitHub discovery

• Enabled lanes: 7
• Configured atomic queries: 26
• Queries executed: 26
• Per-query result cap: 19
• Raw-result budget: 500
• Theoretical maximum raw pointers: 494
• Effective maximum raw pointers: 494
• Raw repository pointers returned: 478
• Unique repositories examined: 459
• Retained after hard rejection: 458
• Hard rejected before enrichment: 1
• Hard rejection reasons: archived: 1

Metadata discovery begins with GitHub repository metadata; bounded read-only enrichment may inspect README and one primary package manifest, but it does not clone or execute repository source trees.

Discovery coverage

• Query families represented: 14
• Lanes represented by candidates: ai-architecture, architectural-comparators, learn-education, platform-engineering, story-game-engine, visual-story, writer-craft
• Candidates matching multiple lanes: 8
• Enrichment shortlist: 25
• README retrieval successes: 25
• README retrieval failures/missing/oversized: 0
• Candidates rescored after enrichment: 25
• Enrichment failure reasons: manifest:missing: 5

Unique candidates per lane

• ai-architecture: 109 unique candidates
• architectural-comparators: 56 unique candidates
• learn-education: 57 unique candidates
• platform-engineering: 55 unique candidates
• story-game-engine: 70 unique candidates
• visual-story: 46 unique candidates
• writer-craft: 75 unique candidates

Query effectiveness

• creative writing agent (writer-craft/story-agents) — current: 18 retained, 0 enriched, 0 Top 5; 1 run(s): 18 retained, 0 enriched, 0 Top 5
• fiction writing agent (writer-craft/story-agents) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• screenwriting AI (writer-craft/screen-world) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• worldbuilding agent (writer-craft/screen-world) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• storyboard AI (visual-story/planning) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• previs tool (visual-story/planning) — current: 8 retained, 0 enriched, 0 Top 5; 1 run(s): 8 retained, 0 enriched, 0 Top 5
• multimodal storytelling (visual-story/multimodal) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• story engine (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• interactive narrative (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• narrative game state (story-game-engine/state-authoring) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• interactive fiction authoring (story-game-engine/state-authoring) — current: 14 retained, 0 enriched, 0 Top 5; 1 run(s): 14 retained, 0 enriched, 0 Top 5
• writing tutor (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• learning feedback agent (learn-education/tutoring) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• reader simulation (learn-education/simulation) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• agent runtime (ai-architecture/runtime-tools) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• agent skills (ai-architecture/runtime-tools) — current: 19 retained, 2 enriched, 1 Top 5; 1 run(s): 19 retained, 2 enriched, 1 Top 5
• tool calling agent (ai-architecture/runtime-tools) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• context orchestration (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• AI agent memory (ai-architecture/context-human) — current: 19 retained, 4 enriched, 1 Top 5; 1 run(s): 19 retained, 4 enriched, 1 Top 5
• human in the loop agent (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• local AI workbench (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• Windows local AI (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 1 run(s): 19 retained, 0 enriched, 0 Top 5
• agent evaluation (platform-engineering/evaluation) — current: 19 retained, 1 enriched, 0 Top 5; 1 run(s): 19 retained, 1 enriched, 0 Top 5
• AI creative workbench (architectural-comparators/creative-systems) — current: 19 retained, 10 enriched, 0 Top 5; 1 run(s): 19 retained, 10 enriched, 0 Top 5
• agentic writing system (architectural-comparators/creative-systems) — current: 19 retained, 3 enriched, 1 Top 5; 1 run(s): 19 retained, 3 enriched, 1 Top 5
• AI production operating system (architectural-comparators/production-systems) — current: 19 retained, 5 enriched, 2 Top 5; 1 run(s): 19 retained, 5 enriched, 2 Top 5

Repository creation age — unique repositories examined

• Created today: 33
• Created yesterday: 11
• Created earlier: 415
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 458
• Seen by Radar yesterday: 0
• Seen by Radar before yesterday: 0

Top 5 Radar history

• 1. openamer/openamer — first seen by Radar today; first seen 2026-09-14
• 2. bendourthe/Nexus-Hub — first seen by Radar today; first seen 2026-09-14
• 3. yaalalabs/agent-kernel — first seen by Radar today; first seen 2026-09-14
• 4. WyattAu/clawdius — first seen by Radar today; first seen 2026-09-14
• 5. xiaohai-uid/cinedrama-os — first seen by Radar today; first seen 2026-09-14

Top repositories for review: 5/5
Meets qualification threshold: 5/5
Below threshold but included for review: 0/5

The daily review queue contains the Top 5 real repositories from today's retained candidates. Qualification status is context only and never blanks the report.

Phase 3 classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Top Repositories for Review

1. openamer/openamer — IMPROVE

The self-improving, self-learning open-source AI agent. Background Computer-Use, A2A Swarm, Vector Memory, Plugin System, Voice, Smart Approvals - 100/100 Score. Windows-native.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 77.31/100
• Discovery Score: 57.01/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI agent memory
• README/package evidence: agent-runtime, education, human-approval, local-ai, mcp-interoperability, memory, multimodal-production, observability, rollback-recovery, skills-plugins, windows-local
• Why PlotPickle should care: Agent Runtime / Context / Provider.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, education, human-approval, local-ai, mcp-interoperability, memory; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Creative Transactions, Evidence, Human Authority, LEARN, Local Story Mode, PPF, Skills, Tools & MCP, Visual Production
• PlotPickle target: Agent Runtime / Context / Provider
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:learn-education, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:education, evidence:human-approval, evidence:local-ai, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:rollback-recovery, evidence:skills-plugins, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

2. bendourthe/Nexus-Hub — IMPROVE

Enterprise-grade skill harness for AI coding assistants -- the upstream catalog for the Nexus desktop app and every major agent (Claude Code, OpenAI Codex, Antigravity, GitHub Copilot, Cursor, GitHub CLI). 336 curated skills, 19 commands, 35 hooks, cross-platform.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 77.13/100
• Discovery Score: 56.83/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agent skills
• README/package evidence: agent-runtime, context-management, education, mcp-interoperability, memory, multimodal-production, observability, retrieval, rollback-recovery, skills-plugins, visual-continuity, windows-local
• Why PlotPickle should care: Agent Runtime / Context / Provider.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, context-management, education, mcp-interoperability, memory, multimodal-production; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Continuity, Creative Transactions, Evidence, LEARN, Local Story Mode, Skills, Tools & MCP, Visual Production
• PlotPickle target: Agent Runtime / Context / Provider
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:learn-education, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:context-management, evidence:education, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:retrieval, evidence:rollback-recovery, evidence:skills-plugins, evidence:visual-continuity, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

3. yaalalabs/agent-kernel — IMPROVE

The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragile glue code. Native support for MCP, A2A. Interface with all mainstream communication channels seamlessly out of the box, production ready from day one.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.91/100
• Discovery Score: 66.76/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval, skills-plugins
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Skills, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:retrieval, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

4. WyattAu/clawdius — IMPROVE

Faster then openclaw, remote agentic coding without worries with wasm sandbox and system bubble wrap to allow reading and prevent writing to system

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.27/100
• Discovery Score: 65.33/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agentic writing system
• README/package evidence: agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability, tool-execution, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Local Story Mode, Responsibility Run, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 1 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:local-ai, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:tool-execution, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

5. xiaohai-uid/cinedrama-os — IMPROVE

Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.00/100
• Discovery Score: 65.10/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Continuity, Local Story Mode, Previs, Skills, Storyboard, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:multimodal-production, evidence:skills-plugins, evidence:storyboard-previs, evidence:visual-continuity, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

[BryanHarrisScripts]

Schedule updated and merged in #2057. The OSS Radar now runs daily at 6:00 AM Eastern using GitHub Actions' native timezone-aware schedule: cron: '0 6 * * *' with timezone: 'America/Toronto'. This keeps the run at 6:00 AM through EST/EDT transitions. workflow_dispatch, least-privilege permissions, concurrency and Radar behavior remain unchanged. Exact-head seven-layer Architecture Verification passed before merge.

[github-actions]

OSS Radar — 2026-09-15

GitHub discovery

• Enabled lanes: 7
• Configured atomic queries: 26
• Queries executed: 26
• Per-query result cap: 19
• Raw-result budget: 500
• Theoretical maximum raw pointers: 494
• Effective maximum raw pointers: 494
• Raw repository pointers returned: 478
• Unique repositories examined: 460
• Retained after hard rejection: 459
• Hard rejected before enrichment: 1
• Hard rejection reasons: archived: 1

Metadata discovery begins with GitHub repository metadata; bounded read-only enrichment may inspect README and one primary package manifest, but it does not clone or execute repository source trees.

Discovery coverage

• Query families represented: 14
• Lanes represented by candidates: ai-architecture, architectural-comparators, learn-education, platform-engineering, story-game-engine, visual-story, writer-craft
• Candidates matching multiple lanes: 8
• Enrichment shortlist: 25
• README retrieval successes: 25
• README retrieval failures/missing/oversized: 0
• Candidates rescored after enrichment: 25
• Enrichment failure reasons: manifest:missing: 7

Unique candidates per lane

• ai-architecture: 110 unique candidates
• architectural-comparators: 56 unique candidates
• learn-education: 56 unique candidates
• platform-engineering: 56 unique candidates
• story-game-engine: 69 unique candidates
• visual-story: 46 unique candidates
• writer-craft: 75 unique candidates

Query effectiveness

• creative writing agent (writer-craft/story-agents) — current: 18 retained, 0 enriched, 0 Top 5; 2 run(s): 36 retained, 0 enriched, 0 Top 5
• fiction writing agent (writer-craft/story-agents) — current: 19 retained, 1 enriched, 0 Top 5; 2 run(s): 38 retained, 2 enriched, 0 Top 5
• screenwriting AI (writer-craft/screen-world) — current: 19 retained, 2 enriched, 2 Top 5; 2 run(s): 38 retained, 3 enriched, 2 Top 5
• worldbuilding agent (writer-craft/screen-world) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• storyboard AI (visual-story/planning) — current: 19 retained, 2 enriched, 2 Top 5; 2 run(s): 38 retained, 3 enriched, 2 Top 5
• previs tool (visual-story/planning) — current: 8 retained, 0 enriched, 0 Top 5; 2 run(s): 16 retained, 0 enriched, 0 Top 5
• multimodal storytelling (visual-story/multimodal) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• story engine (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• interactive narrative (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• narrative game state (story-game-engine/state-authoring) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• interactive fiction authoring (story-game-engine/state-authoring) — current: 14 retained, 0 enriched, 0 Top 5; 2 run(s): 28 retained, 0 enriched, 0 Top 5
• writing tutor (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• learning feedback agent (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 1 enriched, 0 Top 5
• reader simulation (learn-education/simulation) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• agent runtime (ai-architecture/runtime-tools) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 1 enriched, 0 Top 5
• agent skills (ai-architecture/runtime-tools) — current: 19 retained, 2 enriched, 0 Top 5; 2 run(s): 38 retained, 4 enriched, 1 Top 5
• tool calling agent (ai-architecture/runtime-tools) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• context orchestration (ai-architecture/context-human) — current: 19 retained, 1 enriched, 0 Top 5; 2 run(s): 38 retained, 1 enriched, 0 Top 5
• AI agent memory (ai-architecture/context-human) — current: 19 retained, 2 enriched, 0 Top 5; 2 run(s): 38 retained, 6 enriched, 1 Top 5
• human in the loop agent (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• local AI workbench (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 0 enriched, 0 Top 5
• Windows local AI (platform-engineering/local-platform) — current: 19 retained, 1 enriched, 0 Top 5; 2 run(s): 38 retained, 1 enriched, 0 Top 5
• agent evaluation (platform-engineering/evaluation) — current: 19 retained, 0 enriched, 0 Top 5; 2 run(s): 38 retained, 1 enriched, 0 Top 5
• AI creative workbench (architectural-comparators/creative-systems) — current: 19 retained, 11 enriched, 1 Top 5; 2 run(s): 38 retained, 21 enriched, 1 Top 5
• agentic writing system (architectural-comparators/creative-systems) — current: 19 retained, 2 enriched, 0 Top 5; 2 run(s): 38 retained, 5 enriched, 1 Top 5
• AI production operating system (architectural-comparators/production-systems) — current: 19 retained, 5 enriched, 2 Top 5; 2 run(s): 38 retained, 10 enriched, 4 Top 5

Repository creation age — unique repositories examined

• Created today: 29
• Created yesterday: 15
• Created earlier: 416
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 191
• Seen by Radar yesterday: 268
• Seen by Radar before yesterday: 0

Top 5 Radar history

• 1. zenstory-ai/drama-skills — seen by Radar yesterday; first seen 2026-09-14
• 2. yaalalabs/agent-kernel — seen by Radar yesterday; first seen 2026-09-14
• 3. xiaohai-uid/cinedrama-os — seen by Radar yesterday; first seen 2026-09-14
• 4. monarchal-espressoshop454/awesome-ai-storyboarding — seen by Radar yesterday; first seen 2026-09-14
• 5. yeffei/SST — seen by Radar yesterday; first seen 2026-09-14

Top repositories for review: 5/5
Meets qualification threshold: 5/5
Below threshold but included for review: 0/5

The daily review queue contains the Top 5 real repositories from today's retained candidates. Qualification status is context only and never blanks the report.

Phase 3 classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Top Repositories for Review

1. zenstory-ai/drama-skills — LEARN

开源 AI 短剧/漫剧创作 skill 合集：剧本、角色资产、分镜 storyboard、图片/视频提示词、审查，适配 Claude Code 与 Codex | Open-source AI short drama / micro-drama skills for Claude Code & Codex: script, assets, storyboard, image & video prompts, review. MIT.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 76.95/100
• Discovery Score: 62.65/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: screenwriting AI, storyboard AI
• README/package evidence: skills-plugins, windows-local
• Why PlotPickle should care: LEARN / Writers Craft, Visual Story / Storyboard / Previs.
• What can PlotPickle learn from this? Inspect the documented approach to skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Local Story Mode, Skills
• PlotPickle target: LEARN / Writers Craft, Visual Story / Storyboard / Previs
• Craft Module fit: 12
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:skills-plugins, evidence:windows-local, fork:no, lane:visual-story, lane:writer-craft, license:known-open-source, score-stage:enriched

2. yaalalabs/agent-kernel — IMPROVE

The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragile glue code. Native support for MCP, A2A. Interface with all mainstream communication channels seamlessly out of the box, production ready from day one.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.91/100
• Discovery Score: 66.76/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval, skills-plugins
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Skills, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:retrieval, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

3. xiaohai-uid/cinedrama-os — IMPROVE

Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.00/100
• Discovery Score: 65.10/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Continuity, Local Story Mode, Previs, Skills, Storyboard, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 1 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:multimodal-production, evidence:skills-plugins, evidence:storyboard-previs, evidence:visual-continuity, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

4. monarchal-espressoshop454/awesome-ai-storyboarding — LEARN

Find tools for AI storyboarding, pre-production planning, and script-to-visual workflows. A curated list of resources for film directors and producers.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 72.51/100
• Discovery Score: 58.05/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: screenwriting AI, storyboard AI
• README/package evidence: memory, multimodal-production, skills-plugins, storyboard-previs, windows-local
• Why PlotPickle should care: LEARN / Writers Craft, Visual Story / Storyboard / Previs.
• What can PlotPickle learn from this? Inspect the documented approach to memory, multimodal-production, skills-plugins, storyboard-previs, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Local Story Mode, Previs, Skills, Storyboard, Visual Production
• PlotPickle target: LEARN / Writers Craft, Visual Story / Storyboard / Previs
• Craft Module fit: 12
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: CC0-1.0 (review-required)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:memory, evidence:multimodal-production, evidence:skills-plugins, evidence:storyboard-previs, evidence:windows-local, fork:no, lane:visual-story, lane:writer-craft, license:review-required, score-stage:enriched

5. yeffei/SST — IMPROVE

A local-first, extensible creative workbench for AI image generation and editing.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 71.44/100
• Discovery Score: 65.04/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI creative workbench
• README/package evidence: multimodal-production, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to multimodal-production, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Local Story Mode, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: last activity about 67 days ago
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:multimodal-production, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

[github-actions]

OSS Radar — 2026-09-16

GitHub discovery

• Enabled lanes: 7
• Configured atomic queries: 26
• Queries executed: 26
• Per-query result cap: 19
• Raw-result budget: 500
• Theoretical maximum raw pointers: 494
• Effective maximum raw pointers: 494
• Raw repository pointers returned: 478
• Unique repositories examined: 465
• Retained after hard rejection: 464
• Hard rejected before enrichment: 1
• Hard rejection reasons: archived: 1

Metadata discovery begins with GitHub repository metadata; bounded read-only enrichment may inspect README and one primary package manifest, but it does not clone or execute repository source trees.

Discovery coverage

• Query families represented: 14
• Lanes represented by candidates: ai-architecture, architectural-comparators, learn-education, platform-engineering, story-game-engine, visual-story, writer-craft
• Candidates matching multiple lanes: 5
• Enrichment shortlist: 25
• README retrieval successes: 25
• README retrieval failures/missing/oversized: 0
• Candidates rescored after enrichment: 25
• Enrichment failure reasons: manifest:missing: 3, manifest:oversized: 1

Unique candidates per lane

• ai-architecture: 112 unique candidates
• architectural-comparators: 56 unique candidates
• learn-education: 56 unique candidates
• platform-engineering: 56 unique candidates
• story-game-engine: 70 unique candidates
• visual-story: 45 unique candidates
• writer-craft: 75 unique candidates

Query effectiveness

• creative writing agent (writer-craft/story-agents) — current: 18 retained, 1 enriched, 1 Top 5; 3 run(s): 54 retained, 1 enriched, 1 Top 5
• fiction writing agent (writer-craft/story-agents) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 2 enriched, 0 Top 5
• screenwriting AI (writer-craft/screen-world) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 3 enriched, 2 Top 5
• worldbuilding agent (writer-craft/screen-world) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• storyboard AI (visual-story/planning) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 3 enriched, 2 Top 5
• previs tool (visual-story/planning) — current: 8 retained, 0 enriched, 0 Top 5; 3 run(s): 24 retained, 0 enriched, 0 Top 5
• multimodal storytelling (visual-story/multimodal) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• story engine (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• interactive narrative (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• narrative game state (story-game-engine/state-authoring) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• interactive fiction authoring (story-game-engine/state-authoring) — current: 14 retained, 0 enriched, 0 Top 5; 3 run(s): 42 retained, 0 enriched, 0 Top 5
• writing tutor (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• learning feedback agent (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 1 enriched, 0 Top 5
• reader simulation (learn-education/simulation) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• agent runtime (ai-architecture/runtime-tools) — current: 19 retained, 1 enriched, 0 Top 5; 3 run(s): 57 retained, 2 enriched, 0 Top 5
• agent skills (ai-architecture/runtime-tools) — current: 19 retained, 2 enriched, 0 Top 5; 3 run(s): 57 retained, 6 enriched, 1 Top 5
• tool calling agent (ai-architecture/runtime-tools) — current: 19 retained, 3 enriched, 1 Top 5; 3 run(s): 57 retained, 3 enriched, 1 Top 5
• context orchestration (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 1 enriched, 0 Top 5
• AI agent memory (ai-architecture/context-human) — current: 19 retained, 2 enriched, 0 Top 5; 3 run(s): 57 retained, 8 enriched, 1 Top 5
• human in the loop agent (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• local AI workbench (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 0 enriched, 0 Top 5
• Windows local AI (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 1 enriched, 0 Top 5
• agent evaluation (platform-engineering/evaluation) — current: 19 retained, 0 enriched, 0 Top 5; 3 run(s): 57 retained, 1 enriched, 0 Top 5
• AI creative workbench (architectural-comparators/creative-systems) — current: 19 retained, 11 enriched, 0 Top 5; 3 run(s): 57 retained, 32 enriched, 1 Top 5
• agentic writing system (architectural-comparators/creative-systems) — current: 19 retained, 2 enriched, 2 Top 5; 3 run(s): 57 retained, 7 enriched, 3 Top 5
• AI production operating system (architectural-comparators/production-systems) — current: 19 retained, 5 enriched, 2 Top 5; 3 run(s): 57 retained, 15 enriched, 6 Top 5

Repository creation age — unique repositories examined

• Created today: 33
• Created yesterday: 14
• Created earlier: 418
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 182
• Seen by Radar yesterday: 268
• Seen by Radar before yesterday: 14

Top 5 Radar history

• 1. xiaohai-uid/cinedrama-os — seen by Radar yesterday; first seen 2026-09-14
• 2. WyattAu/clawdius — seen by Radar before yesterday; first seen 2026-09-14
• 3. yaalalabs/agent-kernel — seen by Radar yesterday; first seen 2026-09-14
• 4. IBM/mcp-context-forge — first seen by Radar today; first seen 2026-09-16
• 5. Lanerra/saga — first seen by Radar today; first seen 2026-09-16

Top repositories for review: 5/5
Meets qualification threshold: 5/5
Below threshold but included for review: 0/5

The daily review queue contains the Top 5 real repositories from today's retained candidates. Qualification status is context only and never blanks the report.

Phase 3 classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Top Repositories for Review

1. xiaohai-uid/cinedrama-os — IMPROVE

Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.00/100
• Discovery Score: 65.10/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Continuity, Local Story Mode, Previs, Skills, Storyboard, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 2 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:multimodal-production, evidence:skills-plugins, evidence:storyboard-previs, evidence:visual-continuity, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

2. WyattAu/clawdius — IMPROVE

Faster then openclaw, remote agentic coding without worries with wasm sandbox and system bubble wrap to allow reading and prevent writing to system

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.27/100
• Discovery Score: 65.33/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agentic writing system
• README/package evidence: agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability, tool-execution, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Local Story Mode, Responsibility Run, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:local-ai, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:tool-execution, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

3. yaalalabs/agent-kernel — IMPROVE

The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragile glue code. Native support for MCP, A2A. Interface with all mainstream communication channels seamlessly out of the box, production ready from day one.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.91/100
• Discovery Score: 66.76/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval, skills-plugins
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Skills, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:retrieval, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

4. IBM/mcp-context-forge — IMPROVE

An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.88/100
• Discovery Score: 58.21/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: tool calling agent
• README/package evidence: mcp-interoperability, memory, multimodal-production, observability, rollback-recovery, skills-plugins, story-state, tool-execution, windows-local
• Why PlotPickle should care: Agent Runtime / Context / Provider.
• What can PlotPickle learn from this? Inspect the documented approach to mcp-interoperability, memory, multimodal-production, observability, rollback-recovery, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Creative Transactions, Evidence, Local Story Mode, PPF, Responsibility Run, Skills, STORY, Tools & MCP, Visual Production
• PlotPickle target: Agent Runtime / Context / Provider
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence-lane:visual-story, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:rollback-recovery, evidence:skills-plugins, evidence:story-state, evidence:tool-execution, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

5. Lanerra/saga — LEARN

Autonomous, agentic, creative story writing system that incorporates stored embeddings and Knowledge Graphs.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 76.40/100
• Discovery Score: 73.90/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agentic writing system, creative writing agent
• README/package evidence: local-ai, rollback-recovery, skills-plugins, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators, LEARN / Writers Craft.
• What can PlotPickle learn from this? Inspect the documented approach to local-ai, rollback-recovery, skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Creative Transactions, Local Story Mode, Skills
• PlotPickle target: Adjacent Systems / Architectural Comparators, LEARN / Writers Craft
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: Apache-2.0 (known-open-source)
• Activity: active within 1 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:local-ai, evidence:rollback-recovery, evidence:skills-plugins, evidence:windows-local, fork:no, lane:architectural-comparators, lane:writer-craft, license:known-open-source, score-stage:enriched

X-ready public digest

Copy/review this section for public posting. Internal scores, issue IDs, evidence lanes and implementation notes are intentionally omitted.

🎬 Script-to-Screen OSS Radar — 2026-09-16

Open-source work around AI filmmaking keeps expanding. Here are today's projects worth a look:

1. xiaohai-uid/cinedrama-os — STORYBOARD
Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel
Why it matters: Relevant to the handoff from written story intent into storyboard, shot planning or previs.
https://github.com/xiaohai-uid/cinedrama-os

2. WyattAu/clawdius — PRODUCTION
Faster then openclaw, remote agentic coding without worries with wasm sandbox and system bubble wrap to allow reading and prevent writing to system
Why it matters: Useful production reference for moving approved creative work through execution, review or delivery.
https://github.com/WyattAu/clawdius

3. yaalalabs/agent-kernel — PRODUCTION
The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragi…
Why it matters: Useful production reference for moving approved creative work through execution, review or delivery.
https://github.com/yaalalabs/agent-kernel

4. IBM/mcp-context-forge — AGENTS
An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimi…
Why it matters: Useful architecture reference for keeping agents, skills and external tools modular rather than tightly coupled.
https://github.com/IBM/mcp-context-forge

5. Lanerra/saga — WRITING
Autonomous, agentic, creative story writing system that incorporates stored embeddings and Knowledge Graphs.
Why it matters: Useful reference for how open tools support writers, story development or creative learning.
https://github.com/Lanerra/saga

Tracked by the PlotPickle OSS Radar — watching useful ideas, emerging patterns and different approaches across the open-source script-to-screen ecosystem.

#OpenSource #AI #Filmmaking #Screenwriting #GenerativeAI

[github-actions]

OSS Radar — 2026-09-17

GitHub discovery

• Enabled lanes: 7
• Configured atomic queries: 26
• Queries executed: 26
• Per-query result cap: 19
• Raw-result budget: 500
• Theoretical maximum raw pointers: 494
• Effective maximum raw pointers: 494
• Raw repository pointers returned: 478
• Unique repositories examined: 453
• Retained after hard rejection: 452
• Hard rejected before enrichment: 1
• Hard rejection reasons: archived: 1

Metadata discovery begins with GitHub repository metadata; bounded read-only enrichment may inspect README and one primary package manifest, but it does not clone or execute repository source trees.

Discovery coverage

• Query families represented: 14
• Lanes represented by candidates: ai-architecture, architectural-comparators, learn-education, platform-engineering, story-game-engine, visual-story, writer-craft
• Candidates matching multiple lanes: 11
• Enrichment shortlist: 25
• README retrieval successes: 25
• README retrieval failures/missing/oversized: 0
• Candidates rescored after enrichment: 25
• Enrichment failure reasons: manifest:missing: 3, manifest:oversized: 1

Unique candidates per lane

• ai-architecture: 106 unique candidates
• architectural-comparators: 56 unique candidates
• learn-education: 56 unique candidates
• platform-engineering: 55 unique candidates
• story-game-engine: 71 unique candidates
• visual-story: 45 unique candidates
• writer-craft: 75 unique candidates

Query effectiveness

• creative writing agent (writer-craft/story-agents) — current: 18 retained, 1 enriched, 1 Top 5; 4 run(s): 72 retained, 2 enriched, 2 Top 5
• fiction writing agent (writer-craft/story-agents) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 2 enriched, 0 Top 5
• screenwriting AI (writer-craft/screen-world) — current: 19 retained, 1 enriched, 0 Top 5; 4 run(s): 76 retained, 4 enriched, 2 Top 5
• worldbuilding agent (writer-craft/screen-world) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• storyboard AI (visual-story/planning) — current: 19 retained, 1 enriched, 0 Top 5; 4 run(s): 76 retained, 4 enriched, 2 Top 5
• previs tool (visual-story/planning) — current: 8 retained, 0 enriched, 0 Top 5; 4 run(s): 32 retained, 0 enriched, 0 Top 5
• multimodal storytelling (visual-story/multimodal) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• story engine (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• interactive narrative (story-game-engine/engines) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• narrative game state (story-game-engine/state-authoring) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• interactive fiction authoring (story-game-engine/state-authoring) — current: 14 retained, 0 enriched, 0 Top 5; 4 run(s): 56 retained, 0 enriched, 0 Top 5
• writing tutor (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• learning feedback agent (learn-education/tutoring) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 1 enriched, 0 Top 5
• reader simulation (learn-education/simulation) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• agent runtime (ai-architecture/runtime-tools) — current: 19 retained, 2 enriched, 0 Top 5; 4 run(s): 76 retained, 4 enriched, 0 Top 5
• agent skills (ai-architecture/runtime-tools) — current: 19 retained, 1 enriched, 0 Top 5; 4 run(s): 76 retained, 7 enriched, 1 Top 5
• tool calling agent (ai-architecture/runtime-tools) — current: 19 retained, 1 enriched, 1 Top 5; 4 run(s): 76 retained, 4 enriched, 2 Top 5
• context orchestration (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 1 enriched, 0 Top 5
• AI agent memory (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 8 enriched, 1 Top 5
• human in the loop agent (ai-architecture/context-human) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• local AI workbench (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 0 enriched, 0 Top 5
• Windows local AI (platform-engineering/local-platform) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 1 enriched, 0 Top 5
• agent evaluation (platform-engineering/evaluation) — current: 19 retained, 0 enriched, 0 Top 5; 4 run(s): 76 retained, 1 enriched, 0 Top 5
• AI creative workbench (architectural-comparators/creative-systems) — current: 19 retained, 11 enriched, 0 Top 5; 4 run(s): 76 retained, 43 enriched, 1 Top 5
• agentic writing system (architectural-comparators/creative-systems) — current: 19 retained, 3 enriched, 2 Top 5; 4 run(s): 76 retained, 10 enriched, 5 Top 5
• AI production operating system (architectural-comparators/production-systems) — current: 19 retained, 7 enriched, 2 Top 5; 4 run(s): 76 retained, 22 enriched, 8 Top 5

Repository creation age — unique repositories examined

• Created today: 23
• Created yesterday: 15
• Created earlier: 415
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 147
• Seen by Radar yesterday: 275
• Seen by Radar before yesterday: 30

Top 5 Radar history

• 1. xiaohai-uid/cinedrama-os — seen by Radar yesterday; first seen 2026-09-14
• 2. WyattAu/clawdius — seen by Radar yesterday; first seen 2026-09-14
• 3. yaalalabs/agent-kernel — seen by Radar yesterday; first seen 2026-09-14
• 4. IBM/mcp-context-forge — seen by Radar yesterday; first seen 2026-09-16
• 5. Lanerra/saga — seen by Radar yesterday; first seen 2026-09-16

Top repositories for review: 5/5
Meets qualification threshold: 5/5
Below threshold but included for review: 0/5

The daily review queue contains the Top 5 real repositories from today's retained candidates. Qualification status is context only and never blanks the report.

Phase 3 classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Top Repositories for Review

1. xiaohai-uid/cinedrama-os — IMPROVE

Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.00/100
• Discovery Score: 65.10/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to multimodal-production, skills-plugins, storyboard-previs, visual-continuity, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Continuity, Local Story Mode, Previs, Skills, Storyboard, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 3 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:multimodal-production, evidence:skills-plugins, evidence:storyboard-previs, evidence:visual-continuity, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

2. WyattAu/clawdius — IMPROVE

Faster then openclaw, remote agentic coding without worries with wasm sandbox and system bubble wrap to allow reading and prevent writing to system

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.27/100
• Discovery Score: 65.33/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agentic writing system
• README/package evidence: agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability, tool-execution, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, local-ai, mcp-interoperability, memory, multimodal-production, observability; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Local Story Mode, Responsibility Run, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:local-ai, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:tool-execution, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

3. yaalalabs/agent-kernel — IMPROVE

The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragile glue code. Native support for MCP, A2A. Interface with all mainstream communication channels seamlessly out of the box, production ready from day one.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.91/100
• Discovery Score: 66.76/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval, skills-plugins
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, mcp-interoperability, memory, multimodal-production, observability, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Skills, Tools & MCP, Visual Production
• PlotPickle target: Adjacent Systems / Architectural Comparators
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:retrieval, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

4. IBM/mcp-context-forge — IMPROVE

An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.88/100
• Discovery Score: 58.21/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: tool calling agent
• README/package evidence: mcp-interoperability, memory, multimodal-production, observability, rollback-recovery, skills-plugins, story-state, tool-execution, windows-local
• Why PlotPickle should care: Agent Runtime / Context / Provider.
• What can PlotPickle learn from this? Inspect the documented approach to mcp-interoperability, memory, multimodal-production, observability, rollback-recovery, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Creative Transactions, Evidence, Local Story Mode, PPF, Responsibility Run, Skills, STORY, Tools & MCP, Visual Production
• PlotPickle target: Agent Runtime / Context / Provider
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence-lane:visual-story, evidence:mcp-interoperability, evidence:memory, evidence:multimodal-production, evidence:observability, evidence:rollback-recovery, evidence:skills-plugins, evidence:story-state, evidence:tool-execution, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

5. Lanerra/saga — LEARN

Autonomous, agentic, creative story writing system that incorporates stored embeddings and Knowledge Graphs.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 76.40/100
• Discovery Score: 73.90/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agentic writing system, creative writing agent
• README/package evidence: local-ai, rollback-recovery, skills-plugins, windows-local
• Why PlotPickle should care: Adjacent Systems / Architectural Comparators, LEARN / Writers Craft.
• What can PlotPickle learn from this? Inspect the documented approach to local-ai, rollback-recovery, skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Creative Transactions, Local Story Mode, Skills
• PlotPickle target: Adjacent Systems / Architectural Comparators, LEARN / Writers Craft
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: Apache-2.0 (known-open-source)
• Activity: active within 2 days
• Review history: seen in a previous Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:local-ai, evidence:rollback-recovery, evidence:skills-plugins, evidence:windows-local, fork:no, lane:architectural-comparators, lane:writer-craft, license:known-open-source, score-stage:enriched

X-ready public digest

Copy/review this section for public posting. Internal scores, issue IDs, evidence lanes and implementation notes are intentionally omitted.

🎬 Script-to-Screen OSS Radar — 2026-09-17

Open-source work around AI filmmaking keeps expanding. Here are today's projects worth a look:

1. xiaohai-uid/cinedrama-os — STORYBOARD
Industrial-grade AI Anime & Multi-Modal Drama Production Operating System based on Cordis Microkernel
Why it matters: Relevant to the handoff from written story intent into storyboard, shot planning or previs.
https://github.com/xiaohai-uid/cinedrama-os

2. WyattAu/clawdius — PRODUCTION
Faster then openclaw, remote agentic coding without worries with wasm sandbox and system bubble wrap to allow reading and prevent writing to system
Why it matters: Useful production reference for moving approved creative work through execution, review or delivery.
https://github.com/WyattAu/clawdius

3. yaalalabs/agent-kernel — PRODUCTION
The Operating System for Scalable Enterprise AI Agents - Run, orchestrate, and deploy Compliant Enterprise AI Agents at scale across frameworks, without lock-in, rewrites or fragi…
Why it matters: Useful production reference for moving approved creative work through execution, review or delivery.
https://github.com/yaalalabs/agent-kernel

4. IBM/mcp-context-forge — AGENTS
An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimi…
Why it matters: Useful architecture reference for keeping agents, skills and external tools modular rather than tightly coupled.
https://github.com/IBM/mcp-context-forge

5. Lanerra/saga — WRITING
Autonomous, agentic, creative story writing system that incorporates stored embeddings and Knowledge Graphs.
Why it matters: Useful reference for how open tools support writers, story development or creative learning.
https://github.com/Lanerra/saga

Tracked by the PlotPickle OSS Radar — watching useful ideas, emerging patterns and different approaches across the open-source script-to-screen ecosystem.

#OpenSource #AI #Filmmaking #Screenwriting #GenerativeAI

[github-actions]

OSS Radar — 2026-09-18

GitHub discovery

• Enabled discovery lanes: 7
• PlotPickle architecture areas: 7
• Configured atomic queries: 42
• Queries executed: 42
• Per-query result cap: 11
• Raw-result budget: 500
• Theoretical maximum raw pointers: 462
• Effective maximum raw pointers: 462
• Search request pacing: 2250 ms minimum interval
• Search pacing wait: 92250 ms
• Rate-limit retries: 0
• Rate-limit retry wait: 0 ms
• Raw repository pointers returned: 443
• Unique repositories examined: 422
• Retained after hard rejection: 419
• Hard rejected before enrichment: 3
• Hard rejection reasons: archived: 3

Metadata discovery begins with GitHub repository metadata; bounded read-only enrichment may inspect README and one primary package manifest, but it does not clone or execute repository source trees.

Discovery coverage

• Query families represented: 21
• Discovery lanes represented by candidates: ai-architecture, architectural-comparators, learn-education, platform-engineering, story-game-engine, visual-story, writer-craft
• Candidates matching multiple discovery lanes: 8
• Enrichment shortlist: 42
• Lane-balanced enrichment: yes
• README retrieval successes: 41
• README retrieval failures/missing/oversized: 1
• Candidates rescored after enrichment: 41
• Enrichment failure reasons: manifest:missing: 7, manifest:oversized: 1, readme:oversized: 1

Unique candidates per discovery lane

• ai-architecture: 91 unique candidates
• architectural-comparators: 51 unique candidates
• learn-education: 54 unique candidates
• platform-engineering: 64 unique candidates
• story-game-engine: 65 unique candidates
• visual-story: 42 unique candidates
• writer-craft: 63 unique candidates

Query effectiveness

• creative writing agent (writer-craft/story-agents) — current: 10 retained, 1 enriched, 1 provisional selected; 5 run(s): 82 retained, 3 enriched, 3 selected
• fiction writing agent (writer-craft/story-agents) — current: 11 retained, 3 enriched, 2 provisional selected; 5 run(s): 87 retained, 5 enriched, 2 selected
• screenwriting AI (writer-craft/screen-world) — current: 11 retained, 1 enriched, 1 provisional selected; 5 run(s): 87 retained, 5 enriched, 3 selected
• worldbuilding agent (writer-craft/screen-world) — current: 11 retained, 1 enriched, 1 provisional selected; 5 run(s): 87 retained, 1 enriched, 1 selected
• AI authoring workbench (writer-craft/adaptive-authoring) — current: 8 retained, 0 enriched, 0 provisional selected; 1 run(s): 8 retained, 0 enriched, 0 selected
• story production system (writer-craft/adaptive-authoring) — current: 11 retained, 0 enriched, 0 provisional selected; 1 run(s): 11 retained, 0 enriched, 0 selected
• storyboard AI (visual-story/planning) — current: 11 retained, 2 enriched, 1 provisional selected; 5 run(s): 87 retained, 6 enriched, 3 selected
• previs tool (visual-story/planning) — current: 8 retained, 1 enriched, 1 provisional selected; 5 run(s): 40 retained, 1 enriched, 1 selected
• multimodal storytelling (visual-story/multimodal) — current: 11 retained, 1 enriched, 1 provisional selected; 5 run(s): 87 retained, 1 enriched, 1 selected
• generative UI (visual-story/adaptive-interface) — current: 11 retained, 2 enriched, 2 provisional selected; 1 run(s): 11 retained, 2 enriched, 2 selected
• creative workbench UI (visual-story/adaptive-interface) — current: 1 retained, 0 enriched, 0 provisional selected; 1 run(s): 1 retained, 0 enriched, 0 selected
• story engine (story-game-engine/engines) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 87 retained, 0 enriched, 0 selected
• interactive narrative (story-game-engine/engines) — current: 11 retained, 1 enriched, 1 provisional selected; 5 run(s): 87 retained, 1 enriched, 1 selected
• narrative game state (story-game-engine/state-authoring) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 87 retained, 0 enriched, 0 selected
• interactive fiction authoring (story-game-engine/state-authoring) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 67 retained, 0 enriched, 0 selected
• knowledge graph agent (story-game-engine/adaptive-graphs) — current: 11 retained, 3 enriched, 2 provisional selected; 1 run(s): 11 retained, 3 enriched, 2 selected
• context graph (story-game-engine/adaptive-graphs) — current: 11 retained, 2 enriched, 1 provisional selected; 1 run(s): 11 retained, 2 enriched, 1 selected
• writing tutor (learn-education/tutoring) — current: 11 retained, 3 enriched, 1 provisional selected; 5 run(s): 87 retained, 3 enriched, 1 selected
• learning feedback agent (learn-education/tutoring) — current: 10 retained, 1 enriched, 0 provisional selected; 5 run(s): 86 retained, 2 enriched, 0 selected
• reader simulation (learn-education/simulation) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 1 enriched, 0 selected
• human agent collaboration (learn-education/adaptive-human-workflow) — current: 11 retained, 0 enriched, 0 provisional selected; 1 run(s): 11 retained, 0 enriched, 0 selected
• branching workspace AI (learn-education/adaptive-human-workflow) — current: 11 retained, 1 enriched, 1 provisional selected; 1 run(s): 11 retained, 1 enriched, 1 selected
• agent runtime (ai-architecture/runtime-tools) — current: 11 retained, 2 enriched, 1 provisional selected; 5 run(s): 87 retained, 6 enriched, 1 selected
• agent skills (ai-architecture/runtime-tools) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 8 enriched, 1 selected
• tool calling agent (ai-architecture/runtime-tools) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 5 enriched, 2 selected
• context orchestration (ai-architecture/context-human) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 2 enriched, 0 selected
• AI agent memory (ai-architecture/context-human) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 9 enriched, 1 selected
• human in the loop agent (ai-architecture/context-human) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 87 retained, 0 enriched, 0 selected
• coding agent harness (ai-architecture/adaptive-agent-systems) — current: 11 retained, 1 enriched, 0 provisional selected; 1 run(s): 11 retained, 1 enriched, 0 selected
• agent CLI (ai-architecture/adaptive-agent-systems) — current: 11 retained, 2 enriched, 1 provisional selected; 1 run(s): 11 retained, 2 enriched, 1 selected
• MCP gateway (ai-architecture/adaptive-agent-systems) — current: 11 retained, 2 enriched, 0 provisional selected; 1 run(s): 11 retained, 2 enriched, 0 selected
• local AI workbench (platform-engineering/local-platform) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 87 retained, 0 enriched, 0 selected
• Windows local AI (platform-engineering/local-platform) — current: 11 retained, 4 enriched, 0 provisional selected; 5 run(s): 87 retained, 5 enriched, 0 selected
• agent evaluation (platform-engineering/evaluation) — current: 11 retained, 0 enriched, 0 provisional selected; 5 run(s): 87 retained, 1 enriched, 0 selected
• security audit skill (platform-engineering/adaptive-validation) — current: 11 retained, 0 enriched, 0 provisional selected; 1 run(s): 11 retained, 0 enriched, 0 selected
• AI observability (platform-engineering/adaptive-validation) — current: 11 retained, 2 enriched, 0 provisional selected; 1 run(s): 11 retained, 2 enriched, 0 selected
• agent review workflow (platform-engineering/adaptive-validation) — current: 11 retained, 0 enriched, 0 provisional selected; 1 run(s): 11 retained, 0 enriched, 0 selected
• AI creative workbench (architectural-comparators/creative-systems) — current: 11 retained, 3 enriched, 3 provisional selected; 5 run(s): 87 retained, 46 enriched, 4 selected
• agentic writing system (architectural-comparators/creative-systems) — current: 11 retained, 1 enriched, 0 provisional selected; 5 run(s): 87 retained, 11 enriched, 5 selected
• AI production operating system (architectural-comparators/production-systems) — current: 11 retained, 2 enriched, 2 provisional selected; 5 run(s): 87 retained, 24 enriched, 10 selected
• data thread workspace (architectural-comparators/adaptive-comparators) — current: 7 retained, 0 enriched, 0 provisional selected; 1 run(s): 7 retained, 0 enriched, 0 selected
• plugin based AI platform (architectural-comparators/adaptive-comparators) — current: 11 retained, 1 enriched, 0 provisional selected; 1 run(s): 11 retained, 1 enriched, 0 selected

Repository creation age — unique repositories examined

• Created today: 23
• Created yesterday: 10
• Created earlier: 389
• Creation date unavailable: 0

Radar candidate history — retained candidates

• First seen by Radar today: 228
• Seen by Radar yesterday: 167
• Seen by Radar before yesterday: 24

Selected finding history

• 1. thesysdev/openui [What the user sees and interacts with] — first seen by Radar today; first seen 2026-09-18
• 2. MustafaSPN/pictale-showcase [What the user sees and interacts with] — seen by Radar yesterday; first seen 2026-09-14
• 3. Wiplash-ai/glassware [What the user sees and interacts with] — seen by Radar yesterday; first seen 2026-09-14
• 4. danjdewhurst/story-skills [How the experience stays consistent] — seen by Radar yesterday; first seen 2026-09-14
• 5. MuseLinn/MusePi [How the experience stays consistent] — first seen by Radar today; first seen 2026-09-18
• 6. ptonlix/beeweave [How the experience stays consistent] — seen by Radar yesterday; first seen 2026-09-14
• 7. ChrisChen667788/wind-comic [How work moves from idea to finished story] — seen by Radar yesterday; first seen 2026-09-14
• 8. max-sixty/leaf [How work moves from idea to finished story] — first seen by Radar today; first seen 2026-09-18
• 9. 62656456/ai-film-skills [How work moves from idea to finished story] — seen by Radar yesterday; first seen 2026-09-14
• 10. jangtrinh/design-os-3d-blender [How agents, skills and tools work together] — seen by Radar yesterday; first seen 2026-09-17
• 11. LnYo-Cly/ai4j [How agents, skills and tools work together] — first seen by Radar today; first seen 2026-09-18
• 12. leexbo/deepseek-harness-rs [How agents, skills and tools work together] — first seen by Radar today; first seen 2026-09-18
• 13. SaltKing0/ideagraph-live [How story knowledge, canon and evidence stay coherent] — first seen by Radar today; first seen 2026-09-18
• 14. apostrophic-rumination428/Synapse [How story knowledge, canon and evidence stay coherent] — first seen by Radar today; first seen 2026-09-18
• 15. Yuyang16Z/learning-tree [How story knowledge, canon and evidence stay coherent] — first seen by Radar today; first seen 2026-09-18
• 16. ncz-os/mnemos [How models, providers and local/cloud compute plug in] — seen by Radar yesterday; first seen 2026-09-14
• 17. Rinnus/Liagent_OS_V0.1.2 [How models, providers and local/cloud compute plug in] — first seen by Radar today; first seen 2026-09-18
• 18. zenstory-ai/oh-story-claudecode [How models, providers and local/cloud compute plug in] — seen by Radar yesterday; first seen 2026-09-14
• 19. chrismsmr-celcom/agentguard [How PlotPickle stays reliable, secure and maintainable] — first seen by Radar today; first seen 2026-09-18
• 20. ampaduchris/german-b1-tutor [How PlotPickle stays reliable, secure and maintainable] — seen by Radar yesterday; first seen 2026-09-14
• 21. gitlabhq/orbit-knowledge-graph [How PlotPickle stays reliable, secure and maintainable] — first seen by Radar today; first seen 2026-09-18

Architecture findings for review: 21/21
Meets qualification threshold: 21/21
Below threshold but included for review: 0/21
Previously reviewed unchanged candidates suppressed: 2

Today's review queue contains 21 distinct repositories across seven PlotPickle architecture areas. Unchanged previously reviewed repositories are suppressed so the Radar behaves as daily intelligence rather than a persistent leaderboard.

Classifications are deterministic Radar guidance, not adoption decisions. Human review remains authoritative.

Seven PlotPickle Architecture Areas

1. What the user sees and interacts with

Backend alias: Experience Skins

Interfaces, visual surfaces and interaction patterns that make PlotPickle understandable and usable.

1. thesysdev/openui — IMPROVE

The Open Standard for Generative UI

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 83.77/100
• Discovery Score: 53.47/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: generative UI
• README/package evidence: cli-extension, generative-ui, skills-plugins, token-efficiency
• Why PlotPickle should care: What the user sees and interacts with (Experience Skins): Interfaces, visual surfaces and interaction patterns that make PlotPickle understandable and usable.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, generative-ui, skills-plugins, token-efficiency; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Experience Skins, Skills, Tools & MCP
• PlotPickle target: Experience Skins
• Architecture-area evidence: lanes visual-story; concepts generative-ui
• Craft Module fit: 12
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:architectural-comparators, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:cli-extension, evidence:generative-ui, evidence:skills-plugins, evidence:token-efficiency, fork:no, lane:visual-story, license:known-open-source, score-stage:enriched

2. MustafaSPN/pictale-showcase — WATCH

AI storytelling app for children that turns drawings into narrated stories — Flutter client, event-driven Firebase backend, multimodal generation and TTS. Published on the App Store.

• Review status: Meets qualification threshold (55.00/100).
• Enriched PlotPickle Score: 80.26/100
• Discovery Score: 49.20/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: multimodal storytelling
• README/package evidence: creative-writing, education, memory, multimodal-production, rollback-recovery, token-efficiency
• Why PlotPickle should care: What the user sees and interacts with (Experience Skins): Interfaces, visual surfaces and interaction patterns that make PlotPickle understandable and usable.
• What can PlotPickle learn from this? Inspect the documented approach to creative-writing, education, memory, multimodal-production, rollback-recovery, token-efficiency; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Creative Transactions, LEARN, Visual Production, Writer's Craft
• PlotPickle target: Experience Skins
• Architecture-area evidence: lanes visual-story; concepts multimodal-production
• Craft Module fit: 12
• Related Issue fit: none evidenced
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: UNKNOWN (unknown)
• Activity: active within 9 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:learn-education, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence-lane:writer-craft, evidence:creative-writing, evidence:education, evidence:memory, evidence:multimodal-production, evidence:rollback-recovery, evidence:token-efficiency, fork:no, lane:visual-story, license:unknown, score-stage:enriched

3. Wiplash-ai/glassware — IMPROVE

Local-first creative workbench and browser extension with reviewable AI edit plans.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 78.00/100
• Discovery Score: 73.75/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI creative workbench
• README/package evidence: mcp-interoperability, skills-plugins, windows-local
• Why PlotPickle should care: What the user sees and interacts with (Experience Skins): Interfaces, visual surfaces and interaction patterns that make PlotPickle understandable and usable.
• What can PlotPickle learn from this? Inspect the documented approach to mcp-interoperability, skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Local Story Mode, Skills, Tools & MCP
• PlotPickle target: Experience Skins
• Architecture-area evidence: lanes architectural-comparators; concepts windows-local
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 24 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:mcp-interoperability, evidence:skills-plugins, evidence:windows-local, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

2. How the experience stays consistent

Backend alias: Experience Contract

Contracts, schemas, design-system rules and Human-control boundaries that keep every surface behaving consistently.

1. danjdewhurst/story-skills — LEARN

Agent Skills for end-to-end story writing in markdown, packaged as Codex and Claude Code plugins.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 85.19/100
• Discovery Score: 51.08/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: worldbuilding agent
• README/package evidence: cli-extension, memory, skills-plugins, story-state, visual-continuity
• Why PlotPickle should care: How the experience stays consistent (Experience Contract): Contracts, schemas, design-system rules and Human-control boundaries that keep every surface behaving consistently.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, memory, skills-plugins, story-state, visual-continuity; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Continuity, PPF, Skills, STORY, Tools & MCP
• PlotPickle target: Experience Contract
• Architecture-area evidence: lanes none; concepts skills-plugins
• Craft Module fit: 07, 08, 09
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918, LEARN enrichment: independently author new craft teaching from gap analysis #1976
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: MIT (known-open-source)
• Activity: active within 2 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence-lane:visual-story, evidence:cli-extension, evidence:memory, evidence:skills-plugins, evidence:story-state, evidence:visual-continuity, fork:no, lane:writer-craft, license:known-open-source, score-stage:enriched

2. MuseLinn/MusePi — IMPROVE

MusePi — desktop-first AI coding assistant: Electron GUI + TUI/CLI + daemon + desktop companion. Native video understanding, MCP, advisor, memory, multi-domain skills; own update channel.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 73.03/100
• Discovery Score: 71.13/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agent CLI
• README/package evidence: cli-extension, coding-harness, context-management, memory, provider-routing, skills-plugins, ui-contract, windows-local
• Why PlotPickle should care: How the experience stays consistent (Experience Contract): Contracts, schemas, design-system rules and Human-control boundaries that keep every surface behaving consistently.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, coding-harness, context-management, memory, provider-routing, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Experience Contract, Local Story Mode, Skills, Tools & MCP, Verification
• PlotPickle target: Experience Contract
• Architecture-area evidence: lanes ai-architecture; concepts ui-contract, skills-plugins
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:cli-extension, evidence:coding-harness, evidence:context-management, evidence:memory, evidence:provider-routing, evidence:skills-plugins, evidence:ui-contract, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

3. ptonlix/beeweave — IMPROVE

🐝 BeeWeave is an agent-native creation workbench that turns collected material, AI-assisted drafts, and durable markdown knowledge into a reusable creative data flywheel.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.69/100
• Discovery Score: 70.94/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI creative workbench
• README/package evidence: cli-extension, mcp-interoperability, memory, skills-plugins
• Why PlotPickle should care: How the experience stays consistent (Experience Contract): Contracts, schemas, design-system rules and Human-control boundaries that keep every surface behaving consistently.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, mcp-interoperability, memory, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Skills, Tools & MCP
• PlotPickle target: Experience Contract
• Architecture-area evidence: lanes architectural-comparators; concepts skills-plugins
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: last activity about 46 days ago
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:cli-extension, evidence:mcp-interoperability, evidence:memory, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

3. How work moves from idea to finished story

Backend alias: Production Orchestration

Workflows that coordinate planning, generation, review, approval and delivery across long-running creative production.

1. ChrisChen667788/wind-comic — IMPROVE

Multi-agent AI pipeline that turns one line of text into a finished short-form drama: script, cinematic storyboards, character-consistent video. Provider-agnostic (OpenAI/Claude, MJ, Minimax, Veo/Sora, fal, ComfyUI). MIT.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 79.37/100
• Discovery Score: 54.42/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: storyboard AI
• README/package evidence: agent-runtime, api-gateway, local-ai, memory, observability, rollback-recovery, storyboard-previs, visual-continuity
• Why PlotPickle should care: How work moves from idea to finished story (Production Orchestration): Workflows that coordinate planning, generation, review, approval and delivery across long-running creative production.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, api-gateway, local-ai, memory, observability, rollback-recovery; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Continuity, Creative Transactions, Evidence, Local Story Mode, Previs, Storyboard, Tools & MCP
• PlotPickle target: Production Orchestration
• Architecture-area evidence: lanes visual-story; concepts storyboard-previs
• Craft Module fit: 12
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:agent-runtime, evidence:api-gateway, evidence:local-ai, evidence:memory, evidence:observability, evidence:rollback-recovery, evidence:storyboard-previs, evidence:visual-continuity, fork:no, lane:visual-story, license:known-open-source, score-stage:enriched

2. max-sixty/leaf — IMPROVE

Generative UI for your agent

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 81.02/100
• Discovery Score: 50.72/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: generative UI
• README/package evidence: cli-extension, generative-ui, mcp-interoperability, skills-plugins
• Why PlotPickle should care: How work moves from idea to finished story (Production Orchestration): Workflows that coordinate planning, generation, review, approval and delivery across long-running creative production.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, generative-ui, mcp-interoperability, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Experience Skins, Skills, Tools & MCP
• PlotPickle target: Production Orchestration
• Architecture-area evidence: lanes visual-story; concepts none
• Craft Module fit: 12
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:architectural-comparators, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:cli-extension, evidence:generative-ui, evidence:mcp-interoperability, evidence:skills-plugins, fork:no, lane:visual-story, license:known-open-source, score-stage:enriched

3. 62656456/ai-film-skills — LEARN

Open Film Skills: 19 standalone Agent Skills for scripts, directing, storyboards, visual assets, prompts, and AI-video workflows across Codex and Claude Code.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 78.80/100
• Discovery Score: 50.25/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: screenwriting AI
• README/package evidence: cli-extension, skills-plugins, storyboard-previs
• Why PlotPickle should care: How work moves from idea to finished story (Production Orchestration): Workflows that coordinate planning, generation, review, approval and delivery across long-running creative production.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, skills-plugins, storyboard-previs; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Previs, Skills, Storyboard, Tools & MCP
• PlotPickle target: Production Orchestration
• Architecture-area evidence: lanes none; concepts storyboard-previs
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: Apache-2.0 (known-open-source)
• Activity: active within 7 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:visual-story, evidence:cli-extension, evidence:skills-plugins, evidence:storyboard-previs, fork:no, lane:writer-craft, license:known-open-source, score-stage:enriched

4. How agents, skills and tools work together

Backend alias: Agent & Skill Mesh

Agent runtimes, Skills, plugins, CLIs, MCP/tool boundaries and coordination patterns that keep capabilities modular.

1. jangtrinh/design-os-3d-blender — IMPROVE

AI agent operating system for Blender 5.2: bpy knowledge base, AGENT_OK execution contract, 3D-print production gate, worked keyboard and desktop-robot builds.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 79.51/100
• Discovery Score: 75.26/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: agent-runtime, cli-extension, mcp-interoperability, skills-plugins
• Why PlotPickle should care: How agents, skills and tools work together (Agent & Skill Mesh): Agent runtimes, Skills, plugins, CLIs, MCP/tool boundaries and coordination patterns that keep capabilities modular.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, cli-extension, mcp-interoperability, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Skills, Tools & MCP
• PlotPickle target: Agent & Skill Mesh
• Architecture-area evidence: lanes architectural-comparators; concepts agent-runtime, skills-plugins, mcp-interoperability, cli-extension
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:agent-runtime, evidence:cli-extension, evidence:mcp-interoperability, evidence:skills-plugins, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

2. LnYo-Cly/ai4j — IMPROVE

Java 8+ agentic SDK: unified LLM access (OpenAI/Anthropic/DashScope/Doubao/DeepSeek...), Tool Calling, MCP, RAG, Agent Runtime, and a built-in Coding Agent CLI/TUI/ACP.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 73.84/100
• Discovery Score: 72.38/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: tool calling agent
• README/package evidence: agent-runtime, cli-extension, coding-harness, local-ai, mcp-interoperability, retrieval, tool-execution
• Why PlotPickle should care: How agents, skills and tools work together (Agent & Skill Mesh): Agent runtimes, Skills, plugins, CLIs, MCP/tool boundaries and coordination patterns that keep capabilities modular.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, cli-extension, coding-harness, local-ai, mcp-interoperability, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Local Story Mode, Responsibility Run, Skills, Tools & MCP, Verification
• PlotPickle target: Agent & Skill Mesh
• Architecture-area evidence: lanes ai-architecture; concepts agent-runtime, tool-execution, mcp-interoperability, coding-harness, cli-extension
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:agent-runtime, evidence:cli-extension, evidence:coding-harness, evidence:local-ai, evidence:mcp-interoperability, evidence:retrieval, evidence:tool-execution, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

3. leexbo/deepseek-harness-rs — IMPROVE

Native desktop AI coding agent in Rust (GPUI)

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 71.44/100
• Discovery Score: 70.43/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: coding agent harness
• README/package evidence: cli-extension, coding-harness, context-management, mcp-interoperability, skills-plugins, windows-local
• Why PlotPickle should care: How agents, skills and tools work together (Agent & Skill Mesh): Agent runtimes, Skills, plugins, CLIs, MCP/tool boundaries and coordination patterns that keep capabilities modular.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, coding-harness, context-management, mcp-interoperability, skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Local Story Mode, Skills, Tools & MCP, Verification
• PlotPickle target: Agent & Skill Mesh
• Architecture-area evidence: lanes ai-architecture; concepts skills-plugins, mcp-interoperability, coding-harness, cli-extension
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:cli-extension, evidence:coding-harness, evidence:context-management, evidence:mcp-interoperability, evidence:skills-plugins, evidence:windows-local, fork:no, lane:ai-architecture, license:known-open-source, score-stage:enriched

5. How story knowledge, canon and evidence stay coherent

Backend alias: Story / Canon / Evidence

State, memory, context, graphs, retrieval and provenance patterns that preserve story truth and explain where information came from.

1. SaltKing0/ideagraph-live — ADD

Self-growing idea graph: Ingest → Embed → Suggest → Visualize. Generic engine for a persistent knowledge graph backed by your own private git repo.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 74.76/100
• Discovery Score: 52.95/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: knowledge graph agent
• README/package evidence: cli-extension, context-graph, human-approval, mcp-interoperability, memory, retrieval
• Why PlotPickle should care: How story knowledge, canon and evidence stay coherent (Story / Canon / Evidence): State, memory, context, graphs, retrieval and provenance patterns that preserve story truth and explain where information came from.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, context-graph, human-approval, mcp-interoperability, memory, retrieval; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Evidence, Human Authority, PPF, Skills, Tools & MCP
• PlotPickle target: Story / Canon / Evidence
• Architecture-area evidence: lanes story-game-engine; concepts memory, retrieval, context-graph
• Craft Module fit: none evidenced
• Related Issue fit: STORY: The Unwritten — a universal story-game engine #1675
• Recommended review mode: evaluate as a missing capability before opening implementation work
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence:cli-extension, evidence:context-graph, evidence:human-approval, evidence:mcp-interoperability, evidence:memory, evidence:retrieval, fork:no, lane:story-game-engine, license:known-open-source, score-stage:enriched

2. apostrophic-rumination428/Synapse — ADD

Deliver a high-performance knowledge graph combining semantic and text search for efficient AI agent integration using a full MCP interface.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.25/100
• Discovery Score: 73.25/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agent CLI, agent runtime, knowledge graph agent
• README/package evidence: context-graph, windows-local
• Why PlotPickle should care: How story knowledge, canon and evidence stay coherent (Story / Canon / Evidence): State, memory, context, graphs, retrieval and provenance patterns that preserve story truth and explain where information came from.
• What can PlotPickle learn from this? Inspect the documented approach to context-graph, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Evidence, Local Story Mode, PPF
• PlotPickle target: Story / Canon / Evidence
• Architecture-area evidence: lanes story-game-engine, ai-architecture; concepts context-graph
• Craft Module fit: none evidenced
• Related Issue fit: STORY: The Unwritten — a universal story-game engine #1675
• Recommended review mode: evaluate as a missing capability before opening implementation work
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence:context-graph, evidence:windows-local, fork:no, lane:ai-architecture, lane:story-game-engine, license:known-open-source, score-stage:enriched

3. Yuyang16Z/learning-tree — IMPROVE

LearningTree · 学习树 — A local AI learning workspace with branching conversations and bilingual UI.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 75.66/100
• Discovery Score: 51.60/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: branching workspace AI
• README/package evidence: context-management, education, mcp-interoperability, memory, retrieval, windows-local
• Why PlotPickle should care: How story knowledge, canon and evidence stay coherent (Story / Canon / Evidence): State, memory, context, graphs, retrieval and provenance patterns that preserve story truth and explain where information came from.
• What can PlotPickle learn from this? Inspect the documented approach to context-management, education, mcp-interoperability, memory, retrieval, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, LEARN, Local Story Mode, Tools & MCP
• PlotPickle target: Story / Canon / Evidence
• Architecture-area evidence: lanes none; concepts memory, retrieval, context-management
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:learn-education, evidence-lane:platform-engineering, evidence:context-management, evidence:education, evidence:mcp-interoperability, evidence:memory, evidence:retrieval, evidence:windows-local, fork:no, lane:learn-education, license:known-open-source, score-stage:enriched

6. How models, providers and local/cloud compute plug in

Backend alias: Provider Runtime

Model routing, gateways, local/cloud runtimes and interoperability patterns that prevent PlotPickle from locking itself to one provider.

1. ncz-os/mnemos — IMPROVE

Production-grade memory operating system for agentic AI. Interoperates with every major framework via MCP, OpenAI-compatible gateway, and native /v1/* REST — Claude Code, OpenClaw, ZeroClaw, Hermes, MemPalace, LangChain, LlamaIndex, CrewAI, AutoGen, Mem0, Letta, Zep. In production use since December 2025. Apache 2.0.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 79.19/100
• Discovery Score: 72.94/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI production operating system
• README/package evidence: cli-extension, local-ai, mcp-interoperability, memory, observability, provider-routing
• Why PlotPickle should care: How models, providers and local/cloud compute plug in (Provider Runtime): Model routing, gateways, local/cloud runtimes and interoperability patterns that prevent PlotPickle from locking itself to one provider.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, local-ai, mcp-interoperability, memory, observability, provider-routing; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Context Engine, Evidence, Local Story Mode, Skills, Tools & MCP
• PlotPickle target: Provider Runtime
• Architecture-area evidence: lanes architectural-comparators; concepts provider-routing, local-ai, mcp-interoperability
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: Apache-2.0 (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:cli-extension, evidence:local-ai, evidence:mcp-interoperability, evidence:memory, evidence:observability, evidence:provider-routing, fork:no, lane:architectural-comparators, license:known-open-source, score-stage:enriched

2. Rinnus/Liagent_OS_V0.1.2 — IMPROVE

Build a private AI assistant that runs locally with controlled autonomy, combining task management, tool use, and human oversight in one system.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 72.33/100
• Discovery Score: 56.58/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: AI observability
• README/package evidence: agent-runtime, human-approval, local-ai, mcp-interoperability, observability, skills-plugins, windows-local, workflow-orchestration
• Why PlotPickle should care: How models, providers and local/cloud compute plug in (Provider Runtime): Model routing, gateways, local/cloud runtimes and interoperability patterns that prevent PlotPickle from locking itself to one provider.
• What can PlotPickle learn from this? Inspect the documented approach to agent-runtime, human-approval, local-ai, mcp-interoperability, observability, skills-plugins; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, Evidence, Human Authority, Local Story Mode, PPF, Responsibility Run, Skills, Tools & MCP
• PlotPickle target: Provider Runtime
• Architecture-area evidence: lanes platform-engineering; concepts local-ai, windows-local, mcp-interoperability
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:agent-runtime, evidence:human-approval, evidence:local-ai, evidence:mcp-interoperability, evidence:observability, evidence:skills-plugins, evidence:windows-local, evidence:workflow-orchestration, fork:no, lane:platform-engineering, license:known-open-source, score-stage:enriched

3. zenstory-ai/oh-story-claudecode — LEARN

Claude Code / Codex / OpenCode agent skills for writing Chinese web novels (网文): 扫榜、拆文、写作、去AI味、封面全流程，长篇短篇都支持 | 13 skills for novel writing and long-form fiction with file-based continuity tracking. MIT.

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 77.61/100
• Discovery Score: 52.45/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: fiction writing agent
• README/package evidence: cli-extension, mcp-interoperability, skills-plugins, windows-local
• Why PlotPickle should care: How models, providers and local/cloud compute plug in (Provider Runtime): Model routing, gateways, local/cloud runtimes and interoperability patterns that prevent PlotPickle from locking itself to one provider.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, mcp-interoperability, skills-plugins, windows-local; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Local Story Mode, Skills, Tools & MCP
• PlotPickle target: Provider Runtime
• Architecture-area evidence: lanes none; concepts windows-local, mcp-interoperability
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: independently author or adapt learning insight; do not copy source material
• License: MIT (known-open-source)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence:cli-extension, evidence:mcp-interoperability, evidence:skills-plugins, evidence:windows-local, fork:no, lane:writer-craft, license:known-open-source, score-stage:enriched

7. How PlotPickle stays reliable, secure and maintainable

Backend alias: Validation & Operations

Testing, evaluation, review, security, observability, rollback and operational patterns that keep the system trustworthy.

1. chrismsmr-celcom/agentguard — WATCH

Welcome to cerbere-agent guard, originally just agent guard, then rebranded after the mythical hound of the underworld. Our mission? Guarding your AI agents, keeping them in line, and blocking unauthorized actions. Think of it as the ultimate guard dog for AI security and guardrails. 🐕🛡️

• Review status: Meets qualification threshold (55.00/100).
• Enriched PlotPickle Score: 71.09/100
• Discovery Score: 66.13/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: agent runtime, AI observability
• README/package evidence: cli-extension, education, mcp-interoperability, observability, security-audit
• Why PlotPickle should care: How PlotPickle stays reliable, secure and maintainable (Validation & Operations): Testing, evaluation, review, security, observability, rollback and operational patterns that keep the system trustworthy.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, education, mcp-interoperability, observability, security-audit; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Evidence, LEARN, Skills, Tools & MCP, Verification
• PlotPickle target: Validation & Operations
• Architecture-area evidence: lanes platform-engineering, ai-architecture; concepts observability, security-audit
• Craft Module fit: none evidenced
• Related Issue fit: none evidenced
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: NOASSERTION (unknown)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:learn-education, evidence-lane:platform-engineering, evidence:cli-extension, evidence:education, evidence:mcp-interoperability, evidence:observability, evidence:security-audit, fork:no, lane:ai-architecture, lane:platform-engineering, license:unknown, score-stage:enriched

2. ampaduchris/german-b1-tutor — IMPROVE

German B1 writing tutor built entirely by directing a coding agent. Spec-driven, self-verifying, zero hand-written code

• Review status: Meets qualification threshold (65.00/100).
• Enriched PlotPickle Score: 78.15/100
• Discovery Score: 49.60/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: writing tutor
• README/package evidence: coding-harness, story-state
• Why PlotPickle should care: How PlotPickle stays reliable, secure and maintainable (Validation & Operations): Testing, evaluation, review, security, observability, rollback and operational patterns that keep the system trustworthy.
• What can PlotPickle learn from this? Inspect the documented approach to coding-harness, story-state; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Agent Compute, PPF, STORY, Verification
• PlotPickle target: Validation & Operations
• Architecture-area evidence: lanes none; concepts coding-harness
• Craft Module fit: none evidenced
• Related Issue fit: LEARN: 24-course guided journey, open exploration, integrity harness and Agent lenses #1918
• Recommended review mode: adapt architecture or UX idea into an existing PlotPickle system
• License: MIT (known-open-source)
• Activity: active within 6 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence:coding-harness, evidence:story-state, fork:no, lane:learn-education, license:known-open-source, score-stage:enriched

3. gitlabhq/orbit-knowledge-graph — WATCH

Orbit, aka the GitLab Knowledge Graph, is a project that aims to provide a unified context API for AI systems and human users. This project has both a local Knowledge Graph for your code and a backend service for the entire SDLC.

• Review status: Meets qualification threshold (55.00/100).
• Enriched PlotPickle Score: 74.19/100
• Discovery Score: 53.33/100
• Evidence stage: bounded README/package enrichment
• Matched discovery queries: context graph
• README/package evidence: cli-extension, context-graph, mcp-interoperability, observability, security-audit
• Why PlotPickle should care: How PlotPickle stays reliable, secure and maintainable (Validation & Operations): Testing, evaluation, review, security, observability, rollback and operational patterns that keep the system trustworthy.
• What can PlotPickle learn from this? Inspect the documented approach to cli-extension, context-graph, mcp-interoperability, observability, security-audit; bounded README/package evidence is a research signal, not proof of implementation fit.
• PlotPickle evidence target: Context Engine, Evidence, PPF, Skills, Tools & MCP, Verification
• PlotPickle target: Validation & Operations
• Architecture-area evidence: lanes none; concepts observability, security-audit
• Craft Module fit: none evidenced
• Related Issue fit: STORY: The Unwritten — a universal story-game engine #1675
• Recommended review mode: review the repository before deciding whether it merits WATCH or stronger follow-up
• License: NOASSERTION (unknown)
• Activity: active within 0 days
• Review history: first Radar review
• Score evidence: activity:recent, evidence-lane:ai-architecture, evidence-lane:platform-engineering, evidence-lane:story-game-engine, evidence:cli-extension, evidence:context-graph, evidence:mcp-interoperability, evidence:observability, evidence:security-audit, fork:no, lane:story-game-engine, license:unknown, score-stage:enriched

X-ready public digest

Copy/review this section for public posting. Internal scores, issue IDs, evidence lanes and implementation notes are intentionally omitted.

🎬 Story-to-Screen OSS Radar — 2026-09-18

Open-source work around AI filmmaking keeps expanding. Here are today's projects worth a look:

1. thesysdev/openui — STORYBOARD / Experience Skins
The Open Standard for Generative UI
Why it matters: What the user sees and interacts with: Interfaces, visual surfaces and interaction patterns that make PlotPickle understandable and usable.
https://github.com/thesysdev/openui

2. danjdewhurst/story-skills — WRITING / Experience Contract
Agent Skills for end-to-end story writing in markdown, packaged as Codex and Claude Code plugins.
Why it matters: How the experience stays consistent: Contracts, schemas, design-system rules and Human-control boundaries that keep every surface behaving consistently.
https://github.com/danjdewhurst/story-skills

3. ChrisChen667788/wind-comic — STORYBOARD / Production Orchestration
Multi-agent AI pipeline that turns one line of text into a finished short-form drama: script, cinematic storyboards, character-consistent video. Provider-agnostic (OpenAI/Claude,…
Why it matters: How work moves from idea to finished story: Workflows that coordinate planning, generation, review, approval and delivery across long-running creative production.
https://github.com/ChrisChen667788/wind-comic

4. jangtrinh/design-os-3d-blender — PRODUCTION / Agent & Skill Mesh
AI agent operating system for Blender 5.2: bpy knowledge base, AGENT_OK execution contract, 3D-print production gate, worked keyboard and desktop-robot builds.
Why it matters: How agents, skills and tools work together: Agent runtimes, Skills, plugins, CLIs, MCP/tool boundaries and coordination patterns that keep capabilities modular.
https://github.com/jangtrinh/design-os-3d-blender

5. SaltKing0/ideagraph-live — AGENTS / Story / Canon / Evidence
Self-growing idea graph: Ingest → Embed → Suggest → Visualize. Generic engine for a persistent knowledge graph backed by your own private git repo.
Why it matters: How story knowledge, canon and evidence stay coherent: State, memory, context, graphs, retrieval and provenance patterns that preserve story truth and explain where information came from.
https://github.com/SaltKing0/ideagraph-live

6. ncz-os/mnemos — PRODUCTION / Provider Runtime
Production-grade memory operating system for agentic AI. Interoperates with every major framework via MCP, OpenAI-compatible gateway, and native /v1/* REST — Claude Code, OpenClaw…
Why it matters: How models, providers and local/cloud compute plug in: Model routing, gateways, local/cloud runtimes and interoperability patterns that prevent PlotPickle from locking itself to one provider.
https://github.com/ncz-os/mnemos

7. chrismsmr-celcom/agentguard — AGENTS / Validation & Operations
Welcome to cerbere-agent guard, originally just agent guard, then rebranded after the mythical hound of the underworld. Our mission? Guarding your AI agents, keeping them in line,…
Why it matters: How PlotPickle stays reliable, secure and maintainable: Testing, evaluation, review, security, observability, rollback and operational patterns that keep the system trustworthy.
https://github.com/chrismsmr-celcom/agentguard

Presented by PlotPickle — Today’s OSS Radar tracks useful ideas, emerging patterns and different approaches across the open-source script-to-screen ecosystem.

#OpenSource #AI #Filmmaking #Screenwriting #GenerativeAI

---

Presented by PlotPickle — Today’s OSS Radar.