superpowers-reasonix
AGENTS.md101 starsChanged 30 days agoMITView on GitHub
This project ships skills (playbooks) under skills/. They only work if you load them. The rule: before you do ANYTHING non-trivial — before you explore, run bash, write code, or answer — STOP and check the skills index. If a skill might fit, load it. Loading a skill is cheap. Skipping it is the #1 mistake. Process skill FIRST. Action SECOND. Do NOT jump straight to explore or bash or a direct answer. The skill tells you HOW to explore,…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository12 other files it gives its agents.
- superpowers-brainstormingSkill
- superpowers-diagnosing-superpowersSkill
- superpowers-executing-plansSkill
- superpowers-finishing-a-development-branchSkill
- superpowers-receiving-code-reviewSkill
- superpowers-sync-upstreamSkill
- superpowers-systematic-debuggingSkill
- superpowers-test-driven-developmentSkill
- superpowers-using-git-worktreesSkill
- superpowers-verification-before-completionSkill
- superpowers-writing-plansSkill
- superpowers-writing-skillsSkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 7
The rule: before you do ANYTHING non-trivial — before you explore, run bash, write code, or answer — STOP and check the skills index.
- Very specific: line 11
Do NOT jump straight to explore or bash or a direct answer.
- Distinctive: line 38
If you catch yourself about to explore, fix, or answer without loading a skill — STOP and load it.
- Distinctive: line 50
If more than one skill fits, load the process skill first — superpowers-brainstorming, superpowers-systematic-debugging, …
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
Project guidance
Skills — STOP. Load a skill before you act.
This project ships skills (playbooks) under skills/. They only work if you load them.
The rule: before you do ANYTHING non-trivial — before you explore, run bash,
write code, or answer — STOP and check the skills index. If a skill might fit, load it.
Loading a skill is cheap. Skipping it is the #1 mistake. Process skill FIRST. Action SECOND.
Do NOT jump straight to explore or bash or a direct answer. The skill tells you HOW
to explore, debug, build, and verify. So it comes first, every time.
Match the situation:
| If… | Load this FIRST |
| starting a feature, or you have a rough idea | superpowers-brainstorming |
| investigating why a past or current Reasonix session was slow, repetitive, or skipped a skill | superpowers-diagnosing-superpowers |
| a bug, a failing or flaky test, or anything surprising | superpowers-systematic-debugging |
| writing or fixing any code | superpowers-test-driven-development |
| you have a spec for a multi-step task | superpowers-writing-plans |
| executing a written plan in this session | superpowers-executing-plans |
| about to say "done" / "fixed" / "passing" | superpowers-verification-before-completion |
| work is done and tests pass | superpowers-finishing-a-development-branch |
you got code-review feedback (from review or a human) | superpowers-receiving-code-review |
| need an isolated workspace | superpowers-using-git-worktrees |
| making or editing a skill | superpowers-writing-skills |
Load it: run_skill({ name: "<skill-name>", arguments: "<the task>" }).
These skills supplement Reasonix's native tools — they don't replace them. For
dispatching subagents, code review, parallel work, and codebase exploration, use the
native tools directly: task (run a subagent), review (code-review a diff),
wait (join parallel jobs), explore (investigate the codebase). There is no
skill for these — reach for the native tool.
If you catch yourself about to explore, fix, or answer without loading a skill — STOP and load it.
Red flags — these thoughts mean STOP. You are rationalizing.
| You think | Reality |
| "Just a simple question" | Questions are tasks. Check for a skill. |
| "Let me explore / look first" | The skill tells you HOW to explore. Skill first. |
| "I'll just do this one thing first" | Check BEFORE doing anything. |
| "This skill is overkill" | Simple turns complex. Load it. |
| "I already know this" | Knowing ≠ doing. Load the current skill. |
If more than one skill fits, load the process skill first — superpowers-brainstorming, superpowers-systematic-debugging, superpowers-verification-before-completion — then the implementation skill. "Build X" → superpowers-brainstorming first. "Fix this bug" → superpowers-systematic-debugging first.
Priority
- The user's explicit instructions (this file,
REASONIX.md, direct asks) win over everything. - Skills override default behavior where they conflict.
- A user request says WHAT to do, never "skip the skill." "Add X" still means: load the skill, then add X.
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.