Plan Execution Protocol

Important: User-Facing Entry Point

For user-invoked plan execution, use /execute-plan skill.

This skill provides the execution protocols and standards. The execute-plan skill is the primary user-facing interface that wraps this functionality with user intent clarification.

1. Context Discovery (MANDATORY)

Locate and read context files: You MUST locate and read:

• BRIEF.md (project brief)
• PLAN.md (the execution plan)
• ADR.md (architecture decisions, if exists)

Discovery Strategy:

• Follow project-strategy standards for file location.
• Use Glob tool to locate PLAN.md files if needed.

If any required file is missing: Log error and abort.

1.1 Resumability Check (CRITICAL)

State-in-Files Enforcement: Before proceeding with execution, you MUST check for interruption state:

1. Check for HANDOFF.md in current phase directory:

   • Look for .cattoolkit/planning/[phase-name]/HANDOFF.md
   • If found, READ it immediately

2. Resume from Handoff State:

   • Review the "In-Progress" section to understand where execution stopped
   • Review the "Next Actions" section to resume from the correct point
   • DO NOT restart completed tasks listed in the handoff
   • Continue execution from the "Next Actions" specifically

3. Handoff Format Required: When creating HANDOFF.md files, ensure they contain:

   # Handoff - Phase: [Phase Name]
   
   ## Completed Tasks
   - [List of tasks already completed]
   
   ## In-Progress
   - [Current task being worked on, if any]
   
   ## Next Actions
   - [Specific next steps to resume execution]
   
   ## Context
   - [Any relevant context for resuming]
   

4. If no HANDOFF.md exists: Continue with fresh execution as normal.

2. Plan Validation

Validate the injected plan against project-strategy standards:

• YAML frontmatter exists (phase, type: execute, status)
• Objective section is clear
• Context section lists required reading
• Every task has Scope (optional)/Action/Verify/Done
• Success criteria are measurable

3. Strategy Analysis

Analyze each task to identify:

Sequential Chains:

• Task A produces a file that Task B consumes
• Task B cannot start until Task A completes

Parallel Groups:

• Tasks operate on different files with no shared dependencies
• Can execute simultaneously

Async/Background Tasks:

• Long-running tasks (Audit, Test Suite) that block nothing
• Launch with run_in_background: true

4. Delegation

MANDATORY: Delegate to worker subagents with comprehensive context for autonomous execution.

Format: Each worker agent receives natural language instructions with:

• Task: What needs to be done
• Context: Project brief, plan phase, dependencies
• Requirements: Key constraints and success criteria
• Quality Standards: Reference to software-engineering protocols

5. Quality Assurance (CRITICAL)

After ANY subagent reports success, you MUST perform Objective Audit:

1. READ the files they modified - Do NOT trust their report (Read-Back Protocol)
2. Verify against the Plan's "Verify" criteria
3. Update PROJECT STATE - Update PLAN.md status and ROADMAP.md
4. Log verification results

If verification FAILS: Respawn the worker with refined instructions (Correction Loop).

6. Completion

When all tasks pass verification:

1. Update PLAN.md status: status: complete
2. Create SUMMARY.md using project-strategy/assets/templates/summary.md