| name | plan |
| description | Draft a repo-local plan using the plan skill template and optionally save it. |
| metadata | [object Object] |
Plan (Repo)
Draft structured plans for this repository and optionally save them to plan/.
Core rules
- Use
assets/_template.mdas the plan structure and fill every section. - Do not edit code while planning.
- Draft the plan in chat first; ask for confirmation before writing a plan file.
- Save plans to the repo
plan/directory, not~/.codex/plans. - Use the naming pattern:
plan/YYYY-MM-DD_HH-mm-ss-<slug>.md. - The plan must include a matching Issue CSV path:
issues/YYYY-MM-DD_HH-mm-ss-<slug>.csv. - When selecting MCP tools, reference
docs/mcp-tools.mdfor the correctserver:toolnames.
Clarifications
- Ask up to 2 questions if the task is unclear; otherwise state assumptions and proceed.
- If the task continues across turns, re-invoke this skill to keep the rules active.
Plan workflow
- Restate the task and assumptions.
- Draft the plan body in chat (no frontmatter) using the template.
- Ask: "Reply CONFIRM to write the plan file."
- On confirmation, write the plan file with frontmatter via:
python3 .codex/skills/plan/scripts/create_plan.py --task "<short title>" --complexity <simple|medium|complex>- Provide the body via stdin or
--body-file, or use--templatefor a starter.
- If you need to inspect existing plans:
- List:
python3 .codex/skills/plan/scripts/list_plans.py - Read frontmatter:
python3 .codex/skills/plan/scripts/read_plan_frontmatter.py <plan.md>
- List:
Issue CSV
- Generate the Issue CSV after the plan is reviewed/approved.
- Use
assets/_template.csvand fill all required fields. - Follow
issues/README.mdfor column meanings and CSV formatting. - Use
docs/mcp-tools.mdto populate theToolscolumn with validserver:toolnames. - Validate with:
python3 .codex/skills/plan/scripts/validate_issues_csv.py <issues.csv>. - If validation fails, fix and re-run until it passes.