Skip to content
BotAisle
← Back to explore
Free · open source

A planning skill that keeps three markdown files in your project: task_plan.md (phases and checkboxes, the resume point after /clear), findings.md (research notes and decisions) and progress.md (session log and test results). Activated lifecycle hooks re-inject the active plan, so the agent resumes at the current phase after a context reset. It installs across 60+ agents via the Agent Skills standard, with native plugins for Claude Code, Codex CLI, Qoder CLI, Pi, Hermes Agent, OpenCode and DeepSeek Harness.

Highlights

  • The plan is re-read from disk on the next turn after /clear, compaction or a crash
  • Parallel tasks get isolated .planning/ directories with the same three files
  • Optional SHA-256 plan attestation, a parallel-write guard and a gated completion check for long runs
  • Author's evaluation: 96.7% assertion pass rate (29/30) with the skill vs 6.7% without, and 3 of 3 blind A/B wins

Install

Any agent that supports Agent Skills (60+)

npx skills add OthmanAdi/planning-with-files --skill planning-with-files -g

Claude Code (plugin, includes hooks and slash commands)

/plugin marketplace add OthmanAdi/planning-with-files
/plugin install planning-with-files@planning-with-files

Skills

Planning with Files

Creator: Ahmad Othman (@OthmanAdi)

Persistent file-based planning for AI coding agents. The plan lives on disk, so it survives /clear, compaction and crashes.

Price

Free

View source on GitHub ↗

Free listing. You install it directly from the creator's repository.

Official repository

https://github.com/OthmanAdi/planning-with-files

License

MIT

Works with

ClaudeAny

Tags

Agent SkillsSKILL.mdClaude CodeCodex CLICursorPlanningLong-running agents

Security review. Commit de837a9 reviewed on 2026-10-21: no secrets found and no network calls in the shipped scripts. Its hooks run local shell and Python scripts when your agent starts a turn or uses a tool. Security →