#2626th of 43 in Skills, plugins and rules
planning-with-files
Agent skill that keeps task plans in markdown files that survive /clear
- Stars
- 27.4k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
Checked on GitHub.
Overview
Keeps task_plan.md, findings.md and progress.md in the project and uses lifecycle hooks to re-inject the active plan each turn, so a coding agent resumes after /clear, compaction or a crash. A Stop-time completion gate can hold the agent until the plan reports complete. Installs via `npx skills add` for 60+ agents, with native plugins for Claude Code, Codex CLI, Qoder CLI, Pi, Hermes, OpenCode and DeepSeek Harness.
Who it is for: Developers running long tasks in Claude Code, Codex, Cursor or similar agents
Strengths
- Plain markdown plan files on disk, no runtime state or database elsewhere
- Native plugins for several hosts plus the generic Agent Skills install route
- Completion gate holds the agent's stop until the plan reports complete
- Translated variants in 5 languages ship the full script surface
Weaknesses
- Hook behavior and available commands differ per host; Qoder CLI needs Bash and Python 3.10+
- Plan files are gitignored and overwritten by the next task unless you keep them
- Benchmark claims rest on small samples (3 blind A/B comparisons, 30 assertions)
- Completion gate depends on literal English status tokens matched with grep -F
What it needs
- no GPU
- Needs Bash, Python 3.10+ (Qoder CLI), Node.js/npx (installer)
Also in Skills, plugins and rules
See all 43| Rank | Project | Score |
|---|---|---|
| 1 | skillsAgent skills for planning, TDD, debugging and code review | 94 out of 100 |
| 2 | eccAgents, skills, hooks and rules bundle for Claude Code, Codex and other harnesses | 93 out of 100 |
| 3 | ponytailSkill prompt that makes coding agents write less code and add tests | 91 out of 100 |
| #4 | cavemanProxy and skill that cut token use in coding agents | 91 out of 100 |
| #5 | spec-kitCLI and agent skills for spec-driven development, bug fixing and idea assessment | 89 out of 100 |