planning-with-files
Persistent file-based planning for AI coding agents and long-running tasks. Crash-proof markdown plans, session recovery after /clear and compaction, per-turn re-injection against context rot, deterministic completion gate. Manus-style. Install from npm, the Claude Code plugin marketplace, or npx skills. Codex, Cursor, OpenCode, 60+ agents.
- stars
- 26,952
- 30-day movement
- +12040/day
- Related entries
- 61
- Connections
- 3
A skill pack that gives AI coding agents persistent, file-based planning: plans live in markdown files that survive crashes, /clear, and context compaction, with per-turn re-injection and a deterministic completion gate. It follows the Manus-style approach and ships as an agents.md and cursor rules, installable via npm, the Claude Code plugin marketplace, or npx skills.
Long-running agent tasks lose their plan when context resets; this keeps the plan on disk so the agent can recover and stay on track.
Use it to
- Recover agent sessions after /clear or compaction
- Maintain crash-proof markdown plans for long tasks
- Re-inject the plan each turn to fight context rot
- Gate task completion deterministically
- Share planning skills across Codex, Cursor, and other agents
For Developers running AI coding agents on long tasks
- Role
- skill-pack
- Language
- Python
- Licence
- MIT
- Forks
- 2,243
- Open issues
- 9
- Last push
- 2026-09-17
- Latest release
- v1.0.0 · 2026-01-08
- Skills shipped
- 1