specralph
AGENTS.md7 starsChanged 11 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 31,439 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 repository5 other files it gives its agents.
- specs/agents.mdAGENTS.md
- agent-implementationSkill
- agent-spec-creationSkill
- skill-creatorSkill
- spec-creatorSkill
The file
We split this AGENTS.md into the rules it gives an agent
10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 5
Read specs/README.md before any feature work.
- Distinctive: line 8
Verify reality in the codebase before claiming something exists.
- Distinctive: line 10
When Writing specs, NEVER follow Test Driven Development practices.
- Distinctive: line 15
Follow Test Driven Development practices: write failing tests before implementation.
- Very specific: line 20
Run the full Go test suite with make test (includes test/e2e).
- Very specific: line 21
Run only the end-to-end tests with make test-e2e.
- Very specific: line 27
Run from source (no build): make run ARGS='<command> [flags]'.
- Distinctive: line 38
Keep scans deterministic and reproducible.
- Distinctive: line 40
Treat match text as sensitive.
- Distinctive: line 41
When multiple code paths do similar work with small variations, consolidate into shared services with request structs.
≈ “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 Guidelines
Spec-First Workflow
- Read
specs/README.mdbefore any feature work. - Author or update specs with the
spec-creatorskill (.agents/skills/spec-creator) when it is available. - Assume specs describe intent, not implementation.
- Verify reality in the codebase before claiming something exists.
- Implement to spec patterns and data shapes; update specs only when asked.
- When Writing specs, NEVER follow Test Driven Development practices. Write the spec first and stop.
- For programming tasks, follow Test Driven Development practices (see Testing and Quality Gates).
Testing and Quality Gates
- Follow Test Driven Development practices: write failing tests before implementation.
- Local suite:
make quality. - Targeted runs:
make lint|test|test-e2e|test-race|coverage|mutation|security|arch.
- Coverage gate: min 95%.
- Run the full Go test suite with
make test(includestest/e2e). - Run only the end-to-end tests with
make test-e2e. - Execute mutation testing with
make mutationONLY in final stages of the task development. NEVER execute mutation testing during the Test Driven Development process.
Build and Run
- Build the CLI binary:
make build. - Run from source (no build):
make run ARGS='<command> [flags]'. - Get help for available targets:
make help.
Tooling Expectations
- Go version: 1.25 (see
go.mod). - Mutation testing tool:
gremlins. - Lint and security via
golangci-lint,govulncheck,gosec,go-arch-lint,go-fmt.
Implementation Guidance
- Keep scans deterministic and reproducible.
- Skip binary/oversized files per spec; record skipped file stats.
- Treat match text as sensitive; avoid logging it in console.
- When multiple code paths do similar work with small variations, consolidate into shared services with request structs.
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.

