Skip to content

Adjust MAP workflow rules to allow flexible MCP verification#1

Merged
azalio merged 1 commit intomainfrom
codex/conduct-code-review
Oct 21, 2025
Merged

Adjust MAP workflow rules to allow flexible MCP verification#1
azalio merged 1 commit intomainfrom
codex/conduct-code-review

Conversation

@azalio
Copy link
Copy Markdown
Owner

@azalio azalio commented Oct 21, 2025

Summary

  • clarify that Reflector outputs only need to demonstrate that cipher searches informed reasoning without requiring a fixed phrase
  • describe Curator sync_to_cipher checks as conditional on bullets meeting the helpful_count threshold
  • update checklist and testing sections to reflect the more flexible verification guidance

Testing

  • not run (documentation only)

https://chatgpt.com/codex/tasks/task_e_68f713d262e08324b764075e2b91c6f6

Copilot AI review requested due to automatic review settings October 21, 2025 05:15
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR updates the MAP workflow verification rules to be more flexible when checking MCP tool usage by Reflector and Curator agents. Instead of requiring exact output phrases, the documentation now emphasizes evidence-based verification that tools were used correctly.

Key Changes:

  • Replaced rigid phrase-matching requirements with guidance to look for evidence of MCP tool usage
  • Clarified that Curator's sync_to_cipher output is conditional on bullets meeting the helpful_count threshold
  • Updated test verification steps to align with the more flexible approach

Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.

@azalio azalio merged commit 3673006 into main Oct 21, 2025
5 checks passed
azalio added a commit that referenced this pull request Nov 13, 2025
Corrected 'Mistake #1' which had the wrong and correct filenames swapped:
- Wrong mental model should be: 'read .claude/playbook.json' (old, doesn't exist)
- NOT: 'read .claude/playbook.db' (new, correct format)

The mistake was describing the correct approach as wrong.
azalio pushed a commit that referenced this pull request Feb 8, 2026
1. Contract-Based Verification: Replace "adversarial reviewer with 10+
   years experience" role with Protocol-Driven Validation System.
   Monitor now executes a 5-step contract verification sequence:
   parse AAG -> verify Goal -> verify Action -> verify scope -> quality gates.

2. Deterministic REJECT: AAG contract violation is now AUTO-REJECT #1.
   If implementation deviates from Actor -> Action -> Goal, valid=false
   regardless of code aesthetics. The contract IS the specification.

3. Intent Comments Check: Added AUTO-REJECT #9 — missing `# Intent:`
   comments on non-obvious logic blocks, or removal of existing intent
   comments. Ensures next agent in the chain is never "blind".

4. 10-Dimension Cleanup: Replaced "Check dimensions even if early issues
   found" with "Execute validation protocol for each dimension sequentially.
   Do NOT short-circuit." Protocol > vague instruction.

5. Semantic Brackets: Renamed all 16 generic XML tags to Monitor-scoped
   signatures (<adversarial_reviewer> -> <Monitor_Contract_Verification_v2_9>,
   <output_format> -> <Monitor_Output_v2_9>, etc.). Eliminates
   cross-agent tag collision in shared context windows.

https://claude.ai/code/session_01AR3EbNKosxBD5PocKkMSMd
azalio added a commit that referenced this pull request Feb 13, 2026
-7)

- #1-3: Remove Actor "proposals only" section — Actor applies code directly
  with Edit/Write tools, consistent with map-efficient.md prompts
- #4: Rename orchestrator phase 2.7 from APPLY_CHANGES to UPDATE_STATE
  (Actor applies code, 2.7 only updates state tracking)
- #5: Implement check_circuit_breaker command in orchestrator
  (was referenced in map-efficient.md but missing from argparse)
- #6: Replace non-existent map-efficient-step reference with /map-resume
- #7: Fix STEP_ORDER index bug — used [3:] (starts at CHOOSE_MODE)
  instead of index("2.0") (starts at XML_PACKET) for subtask loop reset
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants