m.plan β π€ Review mission.md β m.apply β π€ Review code β [Adjustments] β m.complete
(Handshake #1) (Handshake #2)
- m.plan creates mission.md with INTENT, SCOPE, PLAN, VERIFICATION
- π€ Review & approve the mission before execution (authorize the architecture)
- m.apply executes, polishes, and generates commit message
- π€ Review code and optionally request adjustments (verify the implementation)
- m.complete archives mission and creates git commit
m.debug β π€ Review diagnosis.md β m.plan β m.apply β m.complete
(Investigation) (Fix Planning)
- m.debug investigates the bug and creates diagnosis.md with root cause analysis
- π€ Review diagnosis to understand the problem (evidence-based findings)
- m.plan automatically consumes diagnosis.md to create a targeted fix mission
- m.apply implements the fix with verification
- m.complete archives both diagnosis and fix mission together
.mission/
βββ governance.md # Core principles and workflow rules
βββ backlog.md # Future work and refactoring opportunities
βββ mission.md # Current active mission (auto-generated)
βββ diagnosis.md # Current bug diagnosis (auto-generated)
βββ execution.log # Current mission execution log
βββ completed/ # Archived missions and detailed metrics
βββ paused/ # Temporarily paused missions
βββ libraries/ # Template system (embedded)
# AI-specific prompt directories:
.amazonq/prompts/ # Amazon Q prompts
.claude/commands/ # Claude commands
.kiro/prompts/ # Kiro prompts
.opencode/command/ # OpenCode commands