agent-foundry
AGENTS.md0 starsChanged 8 months agoMITView on GitHub
This repository is the public, reusable abstraction of a private multi-skill system. This repo is meant to be:
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 repository64 other files it gives its agents.
- androidSkill
- aspnet-coreSkill
- best-mindsSkill
- better-iconsSkill
- brainstormingSkill
- chatgpt-appsSkill
- chrome-ext-architectSkill
- compose-skillSkill
- context-compressionSkill
- context-fundamentalsSkill
- creating-skillsSkill
- develop-web-gameSkill
- docSkill
- find-skillsSkill
- fixing-accessibilitySkill
- fixing-motion-performanceSkill
- frontend-designSkill
- gh-address-commentsSkill
- gh-fix-ciSkill
- git-commitSkill
- github-pr-creationSkill
- github-pr-mergeSkill
- github-pr-reviewSkill
- imagegenSkill
- iosSkill
- jupyter-notebookSkill
- linearSkill
- macosSkill
- memory-systemsSkill
- netlify-deploySkill
- notion-knowledge-captureSkill
- notion-meeting-intelligenceSkill
- notion-research-documentationSkill
- notion-spec-to-implementationSkill
- openai-docsSkill
- pdfSkill
- planning-with-filesSkill
- playwrightSkill
- react-best-practicesSkill
- receiving-code-reviewSkill
- render-deploySkill
- requesting-code-reviewSkill
- saas-site-researchSkill
- screenshotSkill
- security-best-practicesSkill
- security-ownership-mapSkill
- security-threat-modelSkill
- sentrySkill
- skill-creatorSkill
- slidesSkill
- soraSkill
- speechSkill
- spreadsheetSkill
- subagent-driven-developmentSkill
- swiftui-expert-skillSkill
- systematic-debuggingSkill
- test-driven-developmentSkill
- theme-factorySkill
- transcribeSkill
- ui-ux-pro-maxSkill
- The first 60 shown.
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 0 are common, 8 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
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 copy personal preferences, private project details, private prompts, or sensitive paths from any private repository.
- Distinctive: line 8
Prefer reusable workflows, generic instructions, measurable evaluation assets, and neutral examples.
- Distinctive: line 9
Keep all public-facing docs bilingual for English and Simplified Chinese when they are part of the main product surface.
- Distinctive: line 10
When adding or changing a public skill, also review:
- Distinctive: line 16
Do not claim an adapter is fully supported unless there is a concrete implementation or a clearly documented manual workflow.
- Very specific: line 20
Keep root instruction files short and stable.
- Distinctive: line 21
Treat skills, rules, hooks, MCP servers, custom agents, and tool configs as privileged surfaces.
- Distinctive: line 36
Prefer concise, product-level language.
- Distinctive: line 38
When describing support for tools, separate:
≈ “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
Public repo guidance
This repository is the public, reusable abstraction of a private multi-skill system.
Core rules
- Do not copy personal preferences, private project details, private prompts, or sensitive paths from any private repository.
- Prefer reusable workflows, generic instructions, measurable evaluation assets, and neutral examples.
- Keep all public-facing docs bilingual for English and Simplified Chinese when they are part of the main product surface.
- When adding or changing a public skill, also review:
catalog/skills-catalog.jsoncatalog/packs.jsondocs/BENCHMARKS.mddocs/SUPPORT_MATRIX.mdREADME.md
- Do not claim an adapter is fully supported unless there is a concrete implementation or a clearly documented manual workflow.
Agent harness updates
- Keep root instruction files short and stable. Put detailed agent harness guidance in
docs/AGENT_HARNESS.md,workflows/, skillreferences/, or scripts. - Treat skills, rules, hooks, MCP servers, custom agents, and tool configs as privileged surfaces. Document scope and run the smallest relevant validation before calling an update ready.
- Public examples must stay neutral: no private paths, personal preferences, private project names, secrets, or host-specific claims without evidence.
Product intent
This repo is meant to be:
- measurable
- installable
- upgradeable
- portable across AI IDEs and CLIs
- community-friendly
Writing style
- Prefer concise, product-level language.
- Favor concrete checklists and repeatable workflows over abstract promises.
- When describing support for tools, separate:
- first-class support
- generated support
- planned support
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.