web-dev-skills / templates
AGENTS.md1 starsChanged 3 months agoMITView on GitHub
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 31,542 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 repository26 other files it gives its agents.
- templates/CLAUDE.mdCLAUDE.md
- website-abuse-preventionSkill
- website-accessibilitySkill
- website-ai-integrationSkill
- website-architectureSkill
- website-authSkill
- website-backend-apiSkill
- website-browser-qaSkill
- website-code-reviewSkill
- website-databaseSkill
- website-design-systemSkill
- website-devopsSkill
- website-ecommerceSkill
- website-frontendSkill
- website-i18nSkill
- website-incident-responseSkill
- website-observabilitySkill
- website-paymentsSkill
- website-performanceSkill
- website-prdSkill
- website-privacySkill
- website-product-discoverySkill
- website-securitySkill
- website-seoSkill
- website-testingSkill
- website-ui-uxSkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 1 is common, 8 are distinctive, and 0 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈8
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
- could be automatic: a setting could enforce it
- Distinctive: line 29
Keep protected authorization server-side and test cross-user/cross-tenant denial.
- Common: line 31
Never print, commit, or expose secret values.
- Distinctive: line 32
Add tests for changed behavior, critical invariants, and negative authorization paths.
- Distinctive: line 34
Keep migrations safe, reviewed, and reversible or explicitly compensated.
- Distinctive: line 35
Keep logs structured and free of sensitive values.
- Distinctive: line 38
Follow the repository threat model and OWASP ASVS requirements.
- Distinctive: line 40
Treat uploaded files, model output, retrieved content, URLs, and webhook payloads as untrusted.
- Distinctive: line 41
Verify webhook signatures, use idempotency, and handle replay/reordering.
- Distinctive: line 42
Use least privilege for tools, cloud roles, tokens, CI jobs, and admin features.
≈ “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
AGENTS.md — Website Engineering Rules
Product and architecture
- Product summary: [ONE PARAGRAPH]
- Primary users and journeys: [LIST]
- Architecture: [LINK/PATH]
- Requirements: [LINK/PATH]
- Threat model: [LINK/PATH]
Repository map
- App/frontend: [PATH]
- Server/API: [PATH]
- Database/migrations: [PATH]
- Shared UI/design system: [PATH]
- Tests: [PATH]
- Infrastructure/deployment: [PATH]
Commands
- Install:
[COMMAND] - Development:
[COMMAND] - Lint:
[COMMAND] - Typecheck:
[COMMAND] - Unit/integration tests:
[COMMAND] - E2E tests:
[COMMAND] - Build:
[COMMAND]
Engineering rules
- Inspect existing patterns before introducing a new abstraction or dependency.
- Keep protected authorization server-side and test cross-user/cross-tenant denial.
- Validate all external input and provider responses at trust boundaries.
- Never print, commit, or expose secret values. Use environment-variable names only.
- Add tests for changed behavior, critical invariants, and negative authorization paths.
- Preserve accessibility, responsive behavior, and error/loading/empty states.
- Keep migrations safe, reviewed, and reversible or explicitly compensated.
- Keep logs structured and free of sensitive values.
Security and abuse rules
- Follow the repository threat model and OWASP ASVS requirements.
- Apply layered rate limits/quotas to authentication, public forms, expensive APIs, uploads, payments, search, and AI actions.
- Treat uploaded files, model output, retrieved content, URLs, and webhook payloads as untrusted.
- Verify webhook signatures, use idempotency, and handle replay/reordering.
- Use least privilege for tools, cloud roles, tokens, CI jobs, and admin features.
Done means
- Acceptance criteria are satisfied.
- Relevant lint, type, tests, and build pass.
- Browser/mobile and accessibility checks are completed for UI work.
- Security/authorization tests pass for protected work.
- The diff is reviewed for regressions, secrets, debug code, and unnecessary dependencies.
- Documentation and operational notes are updated.
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.