plan-on-disk
Skill87 starsChanged 32 days agoMITView on GitHub
Keep the active plan on disk instead of in the context window — task_plan.md, findings.md, progress.md — so it survives /clear, compaction, crashes, and hour ten of a long run. Use when starting multi-step work expected to outlast one sitting or one context window, resuming after a /clear or compaction, when work drifts from the original intent after many tool calls, or when handing a long-running task to an agent or subagent chain. Skip for short tasks that fit in one sitting and never leave the session.
In this repository28 other files it gives its agents.
- adversarial-reviewSkill
- bite-sized-planSkill
- chestertons-fenceSkill
- clarify-the-real-problemSkill
- context-economySkill
- delegate-to-subagentsSkill
- design-for-the-next-readerSkill
- fresh-context-reviewSkill
- honest-artifactsSkill
- interface-contractsSkill
- measure-before-optimizingSkill
- observability-and-instrumentationSkill
- persistent-memorySkill
- premortemSkill
- record-the-whySkill
- root-cause-disciplineSkill
- search-firstSkill
- security-review-by-haloSkill
- security-reviewSkill
- senior-engineer-mindsetSkill
- simplicity-budgetSkill
- surgical-changeSkill
- threat-and-scale-checkSkill
- verifiability-firstSkill
- verify-before-agreeingSkill
- verify-before-claimingSkill
- weigh-tradeoffsSkill
- widen-the-solution-spaceSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Read the file matching the model actually running before applying this skill.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Plan on Disk
The discipline is the same everywhere — three files, recitation over memory, completion as a gate not a feeling — but how it's applied is calibrated per model. Pick the file for the model actually running:
| Model | File |
| Claude Opus 5.5 | opus-5-5.md |
| Claude Fable 5.1 | fable-5-1.md |
| Claude Sonnet 5 | sonnet-5.md |
| Claude Sonnet 5.5 | sonnet-5-5.md |
| Claude Haiku 5.5 | haiku-5-5.md |
| OpenAI GPT-6 Astra | gpt-6-astra.md |
| OpenAI GPT-6 Sol | gpt-6-sol.md |
| OpenAI GPT-6.1 Sol | gpt-6-1-sol.md |
| OpenAI GPT-6 Luna | gpt-6-luna.md |
Read the file matching the model actually running before applying this skill.
Running a model not listed above (an older generation, a different model family entirely)? sonnet-5.md carries the fullest general-purpose version of the underlying procedure — read that one rather than skipping this skill.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.