coding-orchestra / templates
AGENTS.md0 starsChanged 38 days 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 31,774 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 repository7 other files it gives its agents.
- backend-engineeringSkill
- database-api-designSkill
- deployment-readinessSkill
- frontend-engineeringSkill
- production-deliverySkill
- security-auditSkill
- testing-qaSkill
The file
We split this AGENTS.md into the rules it gives an agent
6 rules. About 1 is common, 5 are distinctive, and 0 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈5
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 12
Follow the conventions already in the code (structure, naming, error handling, test style) over your own defaults.
- Distinctive: line 13
Keep changes scoped to what was asked.
- Distinctive: line 14
When goals conflict, the order is: security, data integrity, correctness, compatibility with existing clients, performance, polish.
- Distinctive: line 15
Ask before doing anything that is hard to undo or reaches outside this repository: database migrations and data deletion, changes to …
- Common: line 16
Never print, log or commit secret values, and do not open real env files.
- Distinctive: line 17
After a change, run the project's checks (typecheck, lint, tests, build) and report the real result.
≈ “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
Working agreement
<!--
Optional always-on guidance from Coding Orchestra. Copy the section below into your
project's AGENTS.md (read by Codex and other agents) or CLAUDE.md (read by Claude Code).
It is deliberately short: it covers what an agent cannot infer from the code, which is
how you want risk, scope and reporting handled. Edit it to match your project.
-->How to work in this repository
- Follow the conventions already in the code (structure, naming, error handling, test style) over your own defaults.
- Keep changes scoped to what was asked. Mention unrelated problems you notice; do not fix them in the same change.
- When goals conflict, the order is: security, data integrity, correctness, compatibility with existing clients, performance, polish.
- Ask before doing anything that is hard to undo or reaches outside this repository: database migrations and data deletion, changes to authentication or CORS/CSP behaviour, breaking an API contract, adding infrastructure or a major dependency, deploying, or anything touching production.
- Never print, log or commit secret values, and do not open real env files. Refer to secrets by variable name. A leaked secret needs rotating by a human; removing it from code is not a fix.
- After a change, run the project's checks (typecheck, lint, tests, build) and report the real result. Say what you verified, what you inferred and what you did not test.
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.