asimi
AGENTS.md16 starsChanged 7 days agoMITView on GitHub
Language: Go Write idiomatic Go - Keep it simple, flat, and direct. Edict 1 is reserved for Court Infrastructure operations: - System-level operations (init, bootstrap, project setup) - Infrastructure decisions and precedents This ensures system operations have proper audit trails and prevents edict_id=0 from breaking precedent tracking. Use just <recipe> to manage the project:
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 repository5 other files it gives its agents.
- gh-issue-syncSkill
- write-ritualsSkill
- write-testsSkill
- releaseSkill
- ritualSkill
The file
Show its rules and text
We split this AGENTS.md into the rules it gives an agent
17 rules. About 0 are common, 13 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈13
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
- Distinctive: line 5
Write idiomatic Go: - Keep it simple, flat, and direct.
- Distinctive: line 7
Model friendly file structure: Keep file names specific and module names meaningfull
- Distinctive: line 8
No wrappers: Avoid unnecessary abstractions
- Distinctive: line 9
No build tags: Keep builds simple
- Distinctive: line 10
Short, meaningful names: Follow Go conventions
- Distinctive: line 11
Inline comments: Only for non-trivial code
- Distinctive: line 12
Commit message : start with "fix:", "feat:" or "chore:", follow by short description and end with "e123" for edict 123
- Distinctive: line 13
Few testfiles: don't start a new test file if the there's an exisiting file that can host your tests
- Very specific: line 25
Use just <recipe> to manage the project:
- Distinctive: line 32
default # List all available recipes
- Distinctive: line 44
run # Run with debug logging
- Very specific: line 45
test # Run the test suite, tee-ing output to ./heaven/tests/test-<datetime>.out
- Very specific: line 46
test-atif # ASIMI_PROVIDER) — the test does not force a model.
- Distinctive: line 47
test-coverage # Run tests with coverage
- Distinctive: line 48
test-intent # Run intent/gherkin BDD scenarios
- Distinctive: line 49
test-performance # Run performance benchmarks and guardrail tests
- Very specific: line 50
test-race *packages="./court/" # Run tests under the race detector, tee-ing output to ./heaven/tests/test-race-<datetime>.out
≈ “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
Language & Philosophy
Language: Go
Write idiomatic Go - Keep it simple, flat, and direct.
- Model friendly file structure: Keep file names specific and module names meaningfull
- No wrappers: Avoid unnecessary abstractions
- No build tags: Keep builds simple
- Short, meaningful names: Follow Go conventions
- Inline comments: Only for non-trivial code
- **Commit message **: start with "fix:", "feat:" or "chore:", follow by short description and end with "e123" for edict 123
- Few testfiles: don't start a new test file if the there's an exisiting file that can host your tests
Reserved Edict IDs
Edict 1 is reserved for Court Infrastructure operations:
- System-level operations (init, bootstrap, project setup)
- Infrastructure decisions and precedents
This ensures system operations have proper audit trails and prevents edict_id=0 from breaking precedent tracking.
Common Commands
Use just <recipe> to manage the project:
- bootstrap # Install development tools
- build # Build the binary
- build-sandbox # Build the sandbox container.
- clean # Clean build artifacts
- clean-sandbox # Clean up the sandbox container
- default # List all available recipes
- fmt # Format code
- freeze-logs # Freeze the log files to
./heaven/logs/asimi[-deamon]-<datetime>.log - get-runs # Fetch failed CI runs since the last success, with their entire logs, to heaven/runs.
- get-threats # Run vulnerability scanning, save output to ./heaven/threats/
<tool>-<datetime>.out - hello-harbor *args # Recipe: build the latest asimi for linux, then hand it to Harbor
- init-podman # Init and start the podman machine
- install # Install dependencies
- lint # Run linting
- measure # Measure run_shell_command tool performance
- nvim # Start neovim with the asimi plugin loaded (interactive dev session)
- prune-branches # Li 礼), so resolve those by hand.
- run # Run with debug logging
- test # Run the test suite, tee-ing output to
./heaven/tests/test-<datetime>.out - test-atif # ASIMI_PROVIDER) — the test does not force a model.
- test-coverage # Run tests with coverage
- test-intent # Run intent/gherkin BDD scenarios
- test-performance # Run performance benchmarks and guardrail tests
- test-race *packages="./court/" # Run tests under the race detector, tee-ing output to
./heaven/tests/test-race-<datetime>.out
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.

