dvmctl
AGENTS.md0 starsChanged 32 days agoApache-2.0View on GitHub
Load .agents/skills/public-readme/SKILL.md before changing README.md or a root community file. Keep comments when they explain a constraint, reason, unit, lifecycle, compatibility promise, or API behavior that code alone does not show. Remove narration, stale history, review provenance, and boilerplate. Keep functional JavaScript JSDoc types, licenses, and tooling directives. Public or exported status alone does not require JSDoc. Write tests for a meaningful production failure or an independently specified contract. Exercise the production behavior at the cheapest boundary that catches the…
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 32,671 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 repository3 other files it gives its agents.
- public-readmeSkill
- test-auditSkill
- build-dvmSkill
The file
We split this AGENTS.md into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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
- Very specific: line 9
Before adding, changing, or auditing tests, read .agents/skills/test-audit/SKILL.md.
- Distinctive: line 17
Run the repository verification after the last substantive edit.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 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 guidance
Load .agents/skills/public-readme/SKILL.md before changing README.md or a root community file.
Keep comments when they explain a constraint, reason, unit, lifecycle, compatibility promise, or API behavior that code alone does not show. Remove narration, stale history, review provenance, and boilerplate. Keep functional JavaScript JSDoc types, licenses, and tooling directives. Public or exported status alone does not require JSDoc.
Write tests for a meaningful production failure or an independently specified contract. Exercise the production behavior at the cheapest boundary that catches the failure. Mock external dependencies, not the behavior under test. No tautological tests: do not assert a fixture against itself, language behavior, implementation spelling, prose, or mock output alone. Do not calculate an expected result with the same implementation without an independent contract. Round trips, idempotence and retry/replay properties remain valid when a production defect can violate them.
Before adding, changing, or auditing tests, read .agents/skills/test-audit/SKILL.md.
Source guards are appropriate when a source property is the architecture contract. When adding or changing a guard, include a deliberately invalid example it must reject. Keep release, security, cleanup, recovery, lifecycle, and package-boundary coverage. Before removing a test, identify the retained test that covers each distinct failure, or record that the removed assertion had no meaningful contract. There is no deletion quota. Prose/comment-only and other reversible low-impact edits need no new tests merely because a file changed; required gates still apply. Do not add tests that only pin this guidance. Claims that an observable failure cannot happen require a test reaching that failure; otherwise narrow the claim.
This repository owns the authoritative builder CLI source, behavioral tests, and releases. Author changes directly in src/ and keep the corresponding behavior tests beside the implementation. There is no private-source export or generated source mirror. Consume shared runtime behavior through declared packages and platform behavior through the service contract; do not add dependencies on another repository checkout.
Preserve both source behavior coverage and the node:test package, installed-consumer, and release safeguards. Run npm run test:source -- <path> for focused behavior checks and npm run test:package for artifact checks after building. npm run verify runs both suites with build, lint, and typecheck. Do not add package exports solely to reach code in tests.
Run the repository verification after the last substantive edit. Do not publish from local development. Commit only files belonging to the change, with a public-safe Conventional Commit subject.
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.

