docs: add generic agent workflow for Mission Control assignment (#59)

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.

[itsmiso-ai]

Summary

Adds a comprehensive generic agent workflow document covering the complete lifecycle of agents using Mission Control's assignment layer.

Changes

• New file:docs/agent-workflow.md (237 lines)

Acceptance Criteria Addressed (#59)

Criterion	Status
Document how an agent starts a run	✅ Section 1: "Start a Run" with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2: "Sync Issue State" with POST /api/sync
Document how an agent requests its queue	✅ Section 3: "Request Agent Queue" with GET /api/agents//queue
Document how an agent claims work	✅ Section 4: "Claim Work" with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5: "Report Run Status" with POST /api/agent-runs
Document that GitHub Issues/PRs remain the source of truth	✅ Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Related

Fixes#59

[its-miso]

✅ Automated recommendation: APPROVE

Analysis engine: anthropic/MiniMax-M2.7@https://litellm.jory.dev/v1 (anthropic)

Review: docs: add generic agent workflow for Mission Control assignment (#59)

Recommendation

Approve — This PR adds a comprehensive generic agent workflow document that fully satisfies all acceptance criteria from issue #59.

Change Summary

• File added:docs/agent-workflow.md (237 lines)
• Purpose: Documents the complete lifecycle of agents using Mission Control's assignment layer

Change-by-Change Findings

Section	Finding
Overview	Accurately describes Mission Control as a Postgres-backed cache with GitHub as source of truth
Section 1: Start a Run	Documents POST /api/agent-runs with required/optional fields and 201 response
Section 2: Sync Issue State	Documents POST /api/sync with auth="None required" and failure handling guidance
Section 3: Request Agent Queue	Documents GET /api/agents/<name>/queue with selection priority rules
Section 4: Claim Work	Documents POST /api/issues/claim with normal/force claim behaviors and HTTP status codes (409, 400)
Section 5: Report Run Status	Documents completing a run with metadata
Sections 6-7: Unclaim/Move	Covers optional unclaim and move operations
Source of Truth Rules	Table correctly states GitHub is authoritative; MC is cache; no direct DB writes; no auto-close
Security Constraints	Correctly mandates tokens never be logged, echoed, or persisted
Failure Modes	Documents 5 failure scenarios with recovery guidance; states "MC failures must never crash agent runs"
API Reference Summary	Comprehensive table of 13 endpoints with methods, auth requirements, and purposes
History	Documents creation date and intent to replace scattered notes

Linked Issue Fit

All 9 acceptance criteria from #59 are satisfied:

Criterion	Evidence
Document how an agent starts a run	✅ Section 1 with full POST /api/agent-runs example
Document how an agent syncs issue state	✅ Section 2 with POST /api/sync
Document how an agent requests its queue	✅ Section 3 with GET /api/agents//queue
Document how an agent claims work	✅ Section 4 with POST /api/issues/claim (normal + force)
Document how an agent reports run status	✅ Section 5 with POST /api/agent-runs
Document that GitHub Issues/PRs remain source of truth	✅ "GitHub Issues and PRs remain the source of truth" + Source of Truth Rules table
Document that MC failures should not crash agent runs	✅ Failure Modes section + Critical principle statement
Document that tokens must never be logged	✅ Security Constraints table
Do not mention specific agent names	✅ Uses <agent-name> placeholders throughout

Standards Compliance

AGENTS.md conventions followed:

• ✅ Code Standard Configure Renovate #1: Uses generic patterns (agent/*, <agent-name>) rather than specific agent names
• ✅ Consistent with "OpenClaw Agent Workflow Contract" section in AGENTS.md
• ✅ Follows same endpoint patterns, failure handling guidance, and audit trail requirements

Repository consistency:

• ✅ Endpoints match codebase (/api/sync, /api/agent-runs, /api/issues/claim, /api/issues/unclaim, /api/issues/move)
• ✅ Auth patterns align with existing docs (MISSION_CONTROL_AGENT_TOKEN bearer token)
• ✅ Consistent with docs/openclaw-agent-mc-workflow.md and docs/assignment-conflict-policy.md
• ✅ Table format and documentation style consistent with existing docs

Evidence Provider Findings

No evidence providers configured — no findings to report.

Tool Harness Findings

Tool harness was configured in plan_execute_once mode but no tool requests were executed. No security concerns.

Unknowns / Needs Verification

None — all acceptance criteria are demonstrably met in the diff.

Summary

The PR correctly implements a generic agent workflow document that serves as the single authoritative reference for agents using Mission Control's assignment layer, replacing "section-specific notes scattered across other docs" as noted in the History section.