Plan Writing
Productivityplan-writingPlan Writing is an Agent Skill for productivity published in the a5c-ai/babysitter repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Transform research findings into actionable implementation plans with stakes-based rigor, test-first strategy, and granular task decomposition.
QUICK START
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/a5c-ai/babysitter/blob/HEAD/library/methodologies/rpikit/skills/plan-writing/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/plan-writing, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
- Before implementing any medium or high stakes changes
- When requirements are clear and codebase is understood
Process
- Load research - Find
*-<topic>-research.mdindocs/plans/ - Classify stakes - Low (isolated, reversible), Medium (multiple files), High (architectural)
- Define success criteria - Functional, non-functional, and acceptance criteria
- Decompose tasks - Granular steps with file paths, line references, verification methods
- Plan tests - Test specification as first sub-step per task (test-first)
- Assess risks - Breaking changes, performance, security, dependencies, rollback strategy
- Write plan document -
docs/plans/YYYY-MM-DD-<topic>-plan.md - Approval gate - Human approves, requests changes, or returns to research
Anti-Patterns to Avoid
- Vague task descriptions without specific file references
- Missing verification criteria for any step
- Combining test writing and implementation into single steps
- Planning rigor mismatched to stakes level
- Proceeding without explicit user approval
Tool Use
Invoke via babysitter process: methodologies/rpikit/rpikit-plan