| model | opus |
| name | my-plan |
| description | Create a detailed implementation plan with mechanically verifiable success criteria, phase boundaries, tests-first steps, and explicit non-goals. |
Create Plan
Turn approved research/spec context into an implementation plan that my-implement can execute phase by phase.
Load Rules
Read:
~/.claude/rules/question-policy.md
~/.claude/rules/context-checkpoint.md
~/.claude/rules/tdd-phase.md
Use ~/.agents/rules/ when running through Codex. For complex plans, workflow resumes, or ambiguous scope decisions, read references/protocol-index.md.
Flow
- Resolve task from
$ARGUMENTS, conversation, workflow ledger, research, spec, ticket, or file path.
- Read existing workflow ledger if present; consume linked research/spec artifacts before asking questions.
- Research factual gaps yourself. Pause only for genuine scope, product, or approach decisions.
- Propose implementation approach and boundaries.
- Write ordered phases, each small enough for one executor run.
- For every behavioral phase include RED tests, GREEN changes, allowed paths, verification commands, and success criteria.
- State non-goals and risks explicitly.
- Save plan under
~/.claude/thoughts/shared/plans/.
- Append plan path and assumptions/decisions to workflow ledger when present.
Plan Quality Bar
- Mechanical success criteria, not prose-only “done.”
- One behavior or small unit per phase.
- Tests first for behavioral changes.
- Explicit allowed paths and architectural constraints.
- Clear “what we are not doing.”
Output
Return the plan path, phase summary, decision points, assumptions, and recommended next command.