ai-engineering-skills
AGENTS.md6 starsChanged 3 months agoMITView on GitHub
This repo is mostly Markdown process artifacts plus small standard-library Python scripts. Preserve concise, actionable wording. Do not add unnecessary framework complexity. Do not copy from external skill repos. Keep each skill self-contained. When editing a skill, update related templates if needed. Resolve the task route before loading workflow guidance. See docs/context-isolation.md for the full loading and expansion rules. Run validation after changes: Prefer small diffs. Preserve the MIT license and acknowledgments. Do not add generated marketing fluff. When editing…
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 31,700 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 repository23 other files it gives its agents.
- analyze-miniSkill
- bug-captureSkill
- build-oneSkill
- ceremony-budgetSkill
- checklist-miniSkill
- constitution-liteSkill
- context-checkSkill
- diagnose-loopSkill
- eco-modeSkill
- falsify-specSkill
- grill-with-docs-liteSkill
- handoffSkill
- lean-modeSkill
- mini-specSkill
- operator-profileSkill
- scope-freezeSkill
- ship-miniSkill
- teach-backSkill
- test-miniSkill
- thin-planSkill
- tool-noise-guardSkill
- verify-contractSkill
- workspace-checkpointSkill
The file
Show its rules and text
We split this AGENTS.md into the rules it gives an agent
14 rules. About 0 are common, 13 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈13
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
Do not add unnecessary framework complexity.
- Distinctive: line 9
Do not copy from external skill repos.
- Distinctive: line 11
Keep each skill self-contained.
- Distinctive: line 13
When editing a skill, update related templates if needed.
- Distinctive: line 20
Do not preload the full skill pack merely because it is installed.
- Distinctive: line 23
Treat current project state as more authoritative than templates and examples.
- Distinctive: line 27
Run validation after changes:
- Distinctive: line 39
Do not add generated marketing fluff.
- Very specific: line 41
When editing handoff behavior, preserve Workflow state: active modes, current phase, current loop, next gate, context risk, and active …
- Distinctive: line 45
Every skill must include YAML frontmatter:
- Distinctive: line 50
Every skill must include:
- Distinctive: line 60
Every template must be short and directly usable.
- Distinctive: line 62
Every example must show a concrete use case.
- Distinctive: line 64
Every install path must remain portable across tools.
≈ “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
Agent Instructions
This repo is mostly Markdown process artifacts plus small standard-library Python scripts.
Preserve concise, actionable wording.
Do not add unnecessary framework complexity.
Do not copy from external skill repos.
Keep each skill self-contained.
When editing a skill, update related templates if needed.
Route-first context loading
Resolve the task route before loading workflow guidance.
- Load only the selected skills and directly relevant current-state or project files by default.
- Do not preload the full skill pack merely because it is installed.
- Expand context when the task spans routes, a compatibility seam requires it, or verification exposes missing context.
- Searches and reads remain allowed unless explicitly forbidden. This rule limits the default working set, not discovery.
- Treat current project state as more authoritative than templates and examples.
See docs/context-isolation.md for the full loading and expansion rules.
Run validation after changes:
python -m py_compile scripts/validate_repo.py
python -m py_compile scripts/install_claude_code.py
python scripts/validate_repo.pyPrefer small diffs.
Preserve the MIT license and acknowledgments.
Do not add generated marketing fluff.
When editing handoff behavior, preserve Workflow state: active modes, current phase, current loop, next gate, context risk, and active hypothesis. Do not expand handoff into a transcript summary.
Quality standards
Every skill must include YAML frontmatter:
namedescription
Every skill must include:
- Purpose
- When to use
- Inputs
- Workflow
- Outputs
- Stop conditions
- Anti-patterns
Every template must be short and directly usable.
Every example must show a concrete use case.
Every install path must remain portable across tools.
Claude Code support is first-class.
Codex support is maintained.
Manual install must remain possible because skills are plain folders with SKILL.md files.
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.