AgentHubControlCenter

AgentHubControlCenter is a local-first Personal AI Command Center for managing AI Agent and Skill portfolio projects. V2 upgrades the app from a single portfolio dashboard into an AI Agent Operating System entry point: tools, workflows, useful signals, action center, connectors, and future plugin interface.

Current status: HUB-WCC-INTEGRATION-COMPLETE

Portfolio positioning: this repo is the hub project for showing how 14 local-first AI Agents and Skills can be organized into one safe, inspectable AgentOps workflow system.

Safety mode: public-safe metadata only; no live connector, no credential load, no child project script, and no real action execution.

What It Does

• Loads a local CSV registry of AI Agent and Skill projects.
• Validates registry metadata quality and missing required fields.
• Summarizes portfolio status, showcase coverage, pinned items, public-but-not-pinned projects, and paused/completed projects.
• Checks whether each local project path contains expected public-project files.
• Builds a capability matrix and positioning summary for the portfolio.
• Builds a fixed project matrix view across Finance / Market, Media / OCR / Extraction, Career, News / Signal, SME Automation, Client Delivery / AI Consulting, Knowledge Base, Data Workflow / Insight Engine, Workflow Orchestration / AgentOps, and Control Center / Meta Agent.
• Plans prioritized next actions from validation, health, screenshots, and pin status.
• Provides a local Chinese / English UI toggle for the main command center surfaces.
• Deepens Chinese mode coverage for dynamic display values such as categories, statuses, badges, actions, connectors, next steps, filters, tables, and Agent descriptions while preserving technical IDs and repo names.
• Shows Product Status, Latest Checkpoint, and Manifest Version separately instead of a stale hard-coded stage label.
• Displays local command packs for manual launch, folder open, tests, and git status.
• Generates an enhanced Markdown portfolio report for download or local saved export.
• Builds a Command Center Summary and public-safe showcase asset checklist.
• Builds a Priority Action Summary for paused projects, future commercial candidates, GitHub showcase projects, and future AgentHub integration candidates.
• Shows a V2 Command Overview for what tools are available, what they can do, and what the recommended next action is.
• Defines a standard Agent manifest and contract for future Agent onboarding.
• Scans F:\AIProjects child project folders for agent_manifest.json.
• Validates manifest fields and shows onboarding warnings instead of crashing.
• Merges valid manifest records with data/agent_registry.csv at runtime.
• Imports DataToInsightWorkflowAgent through its AgentHub-compatible manifest and reads its outputs/agenthub_summary.json summary metrics.
• Shows D2I total processed items, high-value signals, medium-value items, low/noise/review items, recommended actions, top routes, and report path.
• Reads WorkflowPackAgent local JSON summaries from F:\AIProjects\WorkflowPackAgent\outputs\ and shows workflow pack count, top workflow packs, safe metadata stats, enriched source agents, and integration status.
• Registers WorkflowCommandCenterAgent as a GitHub-public workflow orchestration / project command spoke with checkpoint WCC-004-GITHUB-PUBLIC-RELEASE-COMPLETE and profile pin status Not pinned.
• Shows source labels: static_registry, local_manifest, and demo_manifest.
• Shows connector readiness without enabling live Gmail, Google Sheets, Notion, Airtable, Telegram, or other external account actions.
• Defines a unified local action schema for every onboarded Agent.
• Shows Action Center metrics for total actions, manual-only actions, display-only actions, future connector actions, required approvals, and blocked actions.
• Shows schema-backed action cards grouped by Agent with risk level, approval flag, expected output, safety note, and manual runbook reference.
• Provides manual runbook and action safety policy docs before any future execution stage.
• Generates copy-ready Codex prompts from Agent manifests, local action metadata, checkpoint context, safety rules, validation requirements, and next recommended actions.
• Scores local/demo Useful Signals from manifest, project status, report, action registry, local JSON/CSV, and manual demo metadata.
• Shows ranked recommendations with usefulness, relevance, urgency, actionability, value, and risk scores without executing any action.
• Simulates future connector readiness for Gmail, Google Sheets, Google Drive, Notion, Airtable, Telegram, GitHub, n8n, Make, and Zapier without connecting live accounts.
• Shows connector permissions, approval gates, risk level, rollback plan, test plan, readiness score, and recommended next step.
• Simulates local demo workflows that chain Agent manifests, Useful Signals, Action Center metadata, Approval Gates, Manual Runbook references, Codex prompts, and summary outputs.
• Shows Approval Gates for display-only, manual-only, template-only, dry-run-only, and blocked workflow steps without enabling real execution.
• Exports public-safe demo workflow reports as Markdown, JSON, and CSV text from local metadata only.
• Provides a Windows desktop launcher path for opening the command center on port 8525.

Why This Project Matters

As the portfolio grows, each project needs a clear operational view: what exists, where it lives, whether the local files are healthy, what it demonstrates, and what needs attention next. This app acts as a PortfolioOps dashboard for a local-first AI workflow ecosystem.

Portfolio Matrix

AgentHubControlCenter is the hub-and-spoke entry point for the current local AI Agent and Skill portfolio.

Project	Category	Role in AgentHub	GitHub status	Backlink status	Manifest status	Public-safe status	Next note
AgentHubControlCenter	Hub / AgentOps Command Center	Main portfolio command center and review hub	Published	Hub project	Valid root manifest	Public-safe metadata only	Keep pinned as the main portfolio hub
WorkflowCommandCenterAgent	Workflow Orchestration / AgentOps / Project Command	Codex instruction, workflow pack, checklist, prompt/rule, and delivery report command spoke	Published: https://github.com/CHENXJC/WorkflowCommandCenterAgent	AgentHub integration documented	Public manifest checked read-only	Public-safe demo workflow command data	Optional profile pin or maintain-showcase
BusinessOpsAgent	SME operations	Business workflow and operations node	Published	Backlink live	Valid manifest	Public-safe	Keep as applied SME workflow example
CareerPilotAgent	Career operations	Career planning and job workflow node	Published	Backlink live	Valid manifest	Public-safe	Keep as practical user-facing workflow example
ClientDeliveryKitAgent	Client delivery / AI automation consulting	Client-facing delivery workflow spoke	Published: https://github.com/CHENXJC/ClientDeliveryKitAgent	Backlink live	Valid published manifest	Public-safe synthetic demo	Optional profile pin / maintain showcase
DataToInsightWorkflowAgent	Data Workflow / Insight Engine	Data-to-insight workflow, scoring, action recommendation, and AgentHub-ready export spoke	Published: https://github.com/CHENXJC/DataToInsightWorkflowAgent	AgentHub integration documented	Valid published manifest + summary JSON	Public-safe synthetic demo	D2I-008 optional profile pin decision
IdeaScoreAgent	Idea validation	Business idea scoring and validation node	Published	Backlink live	Valid manifest	Public-safe	Excluded deploy/report/bat artifacts remain local-only
MarketSenseAgent	Market intelligence	Market watch and local automation node	Local-only non-git	Backlink local-only	Valid local manifest	Public-safe local metadata	Needs separate repo decision before publishing
NewsSignalAgent	News intelligence	News signal analysis node	Published	Backlink live	Valid manifest	Public-safe	Keep as signal analysis example
NextOpsAgent	SME next-action recommendations	Operational next-action recommendation node	Published	Backlink live	Valid manifest	Public-safe	Keep as next-action workflow example
PersonalKnowledgeAgent	Knowledge management	Personal knowledge workflow node	Published	Backlink live	Valid manifest	Public-safe demo boundary	Keep knowledge workflow positioning clear
QuantLabAgent	Quant research	Research, backtesting, and risk analysis node	Published	Backlink live	Valid manifest	Public-safe research/demo boundary	Keep investment disclaimer visible
SocialPainFinderAgent	Opportunity discovery	Pain-point and opportunity discovery node	Published	Backlink live	Valid manifest	Public-safe	Keep as business opportunity discovery example
VideoExtractSkill	Content intelligence	Video/image content extraction node	Local-only non-git	Backlink local-only	Valid local manifest	Public-safe local metadata	Needs separate repo decision before publishing

Screenshots

Command Center Overview

Public-safe command center home view with portfolio metrics, 13 available tools, and first-row Agent cards.

Agent Registry

My Tools / Agent Registry view showing 13 manifest-onboarded local Agents and Skills with source, demo-mode, safe-mode, action, and connector badges.

Action Center

Local Action Schema metrics, policy-safe action table, and Agent-grouped action cards. All actions remain display/manual/template/planned metadata only.

Codex Prompt Generator

Template-only prompt generator with Agent selector, prompt type selector, safety checklist, validation checklist, preview, and copy-ready text area.

Useful Signals

Scored local/demo recommendations for project progress, connector readiness, workflow automation, portfolio improvement, and risk warnings.

Workflow Simulation

My Workflows page showing local workflow simulation metrics and filters. No live connector, credential, child script, or real action is executed.

Connectors

Connector readiness view showing local, link-based, planned, optional, and not-connected surfaces without enabling live account integrations.

Agent Onboarding Metrics

Agent onboarding metrics showing 13 manifests found, 13 valid manifests, 0 invalid manifests, and 0 missing manifests.

Report Export

Demo Workflow Report Export preview for Markdown, JSON, and CSV public-safe text reports under outputs/public_reports/.

Approval Gates

Approval Gates table showing blocked/manual/template-only workflow steps and policy state for high-risk connector ideas.

Core Features

• Product-style Streamlit dashboard
• Personal AI Command Center hero section
• Command Overview page
• My Tools / Agent Registry page
• My Workflows page
• Useful Signals page
• Action Center page
• Connectors page
• Future Plugin Interface page
• Agent Onboarding section
• Manifest discovery and validation
• CSV + manifest runtime registry merge
• Manifest-based Agent onboarding for 13 local Agents and Skills
• HUB-V2-005+ local action schema for 70 metadata-only actions across 13 Agents
• HUB-V2-006 Codex Prompt Generator for copy-ready text prompts
• HUB-V2-007 Useful Signals Engine for scored local recommendations
• HUB-V2-008 Connector Readiness Simulator for design-only connector planning
• HUB-V2-009 Local Workflow Simulation + Approval Gates for demo workflow review
• HUB-V2-010 Demo Workflow Report Export for Markdown / JSON / CSV public-safe reporting
• HUB-V2-011 refreshed 10-screenshot showcase set and sample public-safe report summary
• HUB-V2-012 public showcase release checklist and readiness report
• HUB-V2-013 GitHub showcase update decision, public commit manifest, and public exclusion manifest
• HUB-V2-014 public-safe git commit, push, and live GitHub showcase verification
• HUB-V2-015 Profile Pin / Portfolio Placement Decision with portfolio positioning copy
• HUB-V2-022 bilingual UI toggle and Project Stage sync check
• HUB-V2-023 bilingual UI docs commit and remote sync
• HUB-V2-024 deep Chinese UI coverage check for dynamic AgentHub values
• HUB-V2-025 deep Chinese UI docs commit and remote sync
• Manual runbook references for safe human operation
• Action safety policy for blocked or approval-required action classes
• Reviewed V2 Agent cards with category, source, demo-mode, safe-mode, action, and connector badges
• Six metric cards
• Agent registry table with sidebar filters
• Agent detail panel
• Command pack display
• Registry validation with quality scores
• Health overview cards
• Priority action cards
• Portfolio matrix and capability clusters
• Fixed project matrix view
• Screenshot-ready layout
• Enhanced Markdown report generation and download
• Saved local report export to outputs/
• Command Center report summary
• Priority action summary
• Public showcase readiness section
• Showcase asset checklist
• Standard Agent interface docs and JSON examples
• Codex Prompt Generator docs and text-only prompt policy
• Windows one-click command launcher
• WorkflowPackAgent integration summary for local workflow pack readiness
• WorkflowCommandCenterAgent registry and portfolio matrix integration

Current MVP Status

HUB-WCC-INTEGRATION completes WorkflowCommandCenterAgent integration into AgentHubControlCenter. WCC is tracked as a Workflow Orchestration / AgentOps / Project Command spoke with public repo https://github.com/CHENXJC/WorkflowCommandCenterAgent, checkpoint WCC-004-GITHUB-PUBLIC-RELEASE-COMPLETE, and profile pin status Not pinned. This stage updates AgentHub metadata, docs, and tests only; it does not modify the WCC project, push WCC, execute WCC scripts, or perform profile pinning.

HUB-V2-027 / WPA-004 completes WorkflowPackAgent summary integration. AgentHubControlCenter now reads only local JSON summary files from WorkflowPackAgent and displays workflow pack count, top packs, metadata enriched agents, safe metadata integration status, source metadata stats, next actions, and public-safe integration notes. This is summary-display integration only; it does not execute WorkflowPackAgent scripts, read private files, call APIs, or use external network access.

HUB-V2-024 completes the deep Chinese UI coverage check. Chinese mode now localizes user-facing dynamic values across Agent cards, filters, action cards, connector cards, workflow cards, useful signals, portfolio matrix cards, local tables, and Future Plugin Interface summaries. English mode remains available for public review, and technical proper nouns such as repository names, agent_id, action IDs, connector IDs, schema fields, and file names are preserved.

HUB-V2-022 completed the bilingual UI toggle and stage status sync check. The Streamlit sidebar now offers 中文 and English, defaults to Chinese, and keeps all translations local through agent_hub/ui_i18n.py. The sidebar now separates Product Status, Latest Checkpoint, and Manifest Version through agent_hub/stage_status.py, so the UI no longer displays the stale hard-coded HUB-V2-014 value or treats a cross-project sync checkpoint as the product stage. This is display-only UI polish; it does not execute actions or connect live providers.

CLIENTDELIVERYKIT-010 completes the published status sync for ClientDeliveryKitAgent after live GitHub showcase verification. AgentHubControlCenter now clearly presents itself as the hub-and-spoke entry point for 13 local-first AI Agents and Skills, with 10 published child repos, 2 local-only non-git project notes, valid manifests, public-safe boundaries, and no real execution.

HUB-V2-018 completed the explicit cross-project backlink commit stage. Eight eligible child repos were committed and pushed with exact README.md and agent_manifest.json file paths; MarketSenseAgent and VideoExtractSkill remain local-only non-git directories; IdeaScoreAgent deploy/report/bat artifacts remain untracked and excluded.

HUB-V2-015 completes the GitHub Profile pin / portfolio placement decision. The decision is to strongly recommend pinning AgentHubControlCenter because it is the portfolio hub that best represents the user's AI Agent matrix, workflow automation thinking, safety gates, and AI automation consultant positioning.

HUB-V2-014 completed the authorized public showcase commit/push verification stage on top of the HUB-V2-013 GitHub showcase update decision. It used docs/PUBLIC_COMMIT_FILE_MANIFEST.md as the exact staging source, keeps generated public reports excluded, commits only public-safe project files, and verifies the live GitHub README, screenshot assets, docs, and remote tree after push.

HUB-V2-013 completed a local GitHub showcase update decision on top of the HUB-V2-012 public release check. It classified the working tree into recommended public commit files and intentional exclusions, recorded a suggested staging plan for the explicit commit stage, and confirmed the generated report/output boundaries remain public-safe.

HUB-V2-012 completed the local public showcase release check on top of the V2 action schema, manual runbook, Codex Prompt Generator, Useful Signals Engine, Connector Readiness Simulator, Local Workflow Simulation, Approval Gates, and Demo Workflow Report Export layers. It verified the 10-image screenshot set, sample public-safe report summary, manifest/contract JSON, launcher, .gitignore boundaries, policy counts, and release-readiness docs for GitHub/portfolio review.

The My Workflows page now includes workflow simulation metrics, workflow cards, Approval Gates, workflow-generated Useful Signals, and Demo Workflow Report Export previews/downloads. Each workflow remains local_simulation_only_no_live_connector_no_real_action_no_credentials, and report export remains public_safe_demo_report_metadata_only_no_execution.

The Connectors page still includes readiness metrics, risk/status/provider filters, connector cards, readiness table, and connector-generated Useful Signals. Each connector remains not_connected and design_only_readiness_simulation_no_live_connection.

The V2 stack now sits on top of the completed HUB-V2-001 / HUB-V2-001A command center and launcher work, the HUB-V2-002 safe discovery layer, the HUB-V2-003 child manifests, and the HUB-V2-004 registry/screenshot review.

It does not call OpenAI APIs, does not create OAuth flows, does not create credentials, does not execute child project scripts, does not execute action command templates, does not connect external accounts, and does not auto-send generated prompts or recommendations anywhere. GitHub push is only performed in the explicit V2-014 release stage authorized by the user, without changing git remotes or using force push.

Current local scan result:

• Total projects scanned: 12
• Manifests found: 12
• Valid manifests: 12
• Invalid manifests: 0
• Missing manifests: 0
• Imported agents: 12
• Duplicate agent IDs: 9 static registry overrides

Current local action result:

• Total actions: 66
• Manual-only actions: 19
• Display-only actions: 34
• Future connector actions: 0
• Requires approval: 3
• Blocked actions: 0
• Policy violations: 0

Current useful signal result:

• Demo/local useful signals: 20 including 3 connector-generated signals and 3 workflow-generated signals
• Execution policy: display_only_text_recommendation_no_execution
• Real action execution: 0

Current connector readiness result:

• Demo connectors: 14
• Design-only connectors: 1
• Ready for demo: 2
• Needs review: 2
• Blocked until approved: 6
• High-risk connectors: 10
• Average readiness score: 44.9
• Live connectors connected: 0
• Connector policy violations: 0

Current local workflow simulation result:

• Demo workflows: 4
• Workflow-generated Useful Signals: 3
• Live connectors connected: 0
• Real workflow execution: 0
• Workflow policy violations: 0
• Approval gate policy violations: 0

Current demo workflow report export result:

• Available report sections: 7
• Export formats: Markdown, JSON, CSV
• Output directory: outputs/public_reports/
• Report export policy: public_safe_demo_report_metadata_only_no_execution
• Report export policy violations: 0

Current showcase refresh result:

• Canonical screenshots: 10 under docs/images/
• Sample report summary: docs/SAMPLE_DEMO_WORKFLOW_REPORT_SUMMARY.md
• Screenshot public-safe review: no .env, secret, token, password, API key, credential, private output, live connector data, or real execution visible
• Public-safe report source checked: latest Markdown/JSON/CSV files under outputs/public_reports/

Current release readiness result:

• Public release checklist: docs/PUBLIC_RELEASE_CHECKLIST.md
• V2 release readiness report: docs/V2_RELEASE_READINESS_REPORT.md
• README/docs/screenshots consistency: ready
• Manifest/contract JSON validation: ready
• Public reports boundary: generated report files remain ignored; .gitkeep is the only intended tracked file under outputs/public_reports/
• Release recommendation: ready for a separate explicit commit/push decision

Registered Agents

• AgentHubControlCenter
• WorkflowCommandCenterAgent
• ClientDeliveryKitAgent
• DataToInsightWorkflowAgent
• VideoExtractSkill
• MarketSenseAgent
• QuantLabAgent
• SocialPainFinderAgent
• CareerPilotAgent
• NewsSignalAgent
• PersonalKnowledgeAgent
• BusinessOpsAgent
• NextOpsAgent
• IdeaScoreAgent

Tech Stack

• Python 3.11+
• Streamlit
• pandas
• pytest

How To Run

cd F:\AIProjects\AgentHubControlCenter
python -m pip install -r requirements.txt
streamlit run app.py

Desktop launcher:

cd F:\AIProjects\AgentHubControlCenter
.\launch_command_center.cmd

Create desktop shortcut:

cd F:\AIProjects\AgentHubControlCenter
powershell -ExecutionPolicy Bypass -File .\scripts\create_desktop_shortcut.ps1

Run tests:

cd F:\AIProjects\AgentHubControlCenter
python -m pytest
python -m compileall .

Safety Boundary

This dashboard is for local portfolio management and workflow planning only. It does not execute external actions or access private credentials.

The project intentionally avoids .env creation, API integrations, token reads, credential storage, remote initialization, remote modification, and force push. HUB-V2-014 performs only the explicitly authorized public-safe git staging, commit, normal push, and live showcase verification. Connectors, workflows, approval gates, reports, screenshots, and sample summaries are shown as local/demo/planned/readiness/simulation/export/showcase surfaces only. Manifest discovery only reads agent_manifest.json files from immediate child project folders, and child project actions remain display/manual/template/planned metadata. Generated Codex prompts and Useful Signals recommendations are text-only planning artifacts. Connector readiness records are design-only metadata and do not load credentials or call provider APIs. Workflow simulations and Approval Gates are metadata-only and do not execute real actions. Demo Workflow Report Export writes only public-safe text under outputs/public_reports/ when local file export is used, and does not write outputs/private/. The showcase refresh only updates public-safe docs and screenshots. The bilingual UI toggle uses a local dictionary only and does not call translation APIs.

Roadmap

• GitHub Public Release: complete.
• HUB-V2-001 Unified Command Center Entry + Desktop Launcher: complete.
• HUB-V2-001A Desktop Launcher Runtime Fix: complete.
• HUB-V2-002 Manifest Import + Agent Onboarding Flow: complete.
• HUB-V2-003 Child Agent Manifest Templates: complete.
• HUB-V2-004 V2 Onboarding Screenshot Refresh + Registry Review: complete.
• HUB-V2-005 Local Action Schema + Manual Runbook: complete.
• HUB-V2-006 Codex Prompt Generator: complete.
• HUB-V2-007 Useful Signals Engine: complete.
• HUB-V2-008 Connector Readiness Simulator: complete.
• HUB-V2-009 Local Workflow Simulation + Approval Gates: complete.
• HUB-V2-010 Demo Workflow Report Export: complete.
• HUB-V2-011 Report Showcase / Screenshot Refresh: complete.
• HUB-V2-012 Public Showcase Release Check: complete.
• HUB-V2-013 GitHub Showcase Update Decision: complete.
• HUB-V2-014 Git Commit / Push / Live Showcase Verification: complete.
• HUB-V2-015 Profile Pin / Portfolio Placement Decision: complete.
• HUB-V2-016 Cross-Project Backlink Plan: complete.
• HUB-V2-017 Cross-Project Backlink Commit Decision: complete.
• HUB-V2-018 Cross-Project Backlink Explicit Commit: complete.
• HUB-V2-019 Portfolio Matrix Final Review: complete.
• HUB-V2-020 Project Pause and Next Portfolio Gap Decision: complete.
• HUB-V2-021 AgentHub Pause Docs Commit: complete.
• HUB-V2-022 Bilingual UI Toggle and Stage Sync Check: complete.
• HUB-V2-023 Bilingual UI Docs Commit: complete.
• HUB-V2-024 Deep Chinese UI Coverage Check: complete.
• HUB-V2-025 Deep Chinese UI Docs Commit: complete.