free-kimi-code
AGENTS.md3 starsChanged 53 days agoMITView on GitHub
Route: implementation/UI → lazy-developer; failure → lazy-debug; verification → lazy-test; review/audit → lazy-review. Inspect only the relevant scope. Preserve working behavior. Never invent tests/results. Standalone deliverables use LAZYDEV_ARTIFACT_DIR; on Termux/PRoot Debian/Ubuntu this is /storage/emulated/0/lazydevfile, while Windows/Linux/macOS use the user-home lazydevfile directory. Native CLI data stays in each tool’s normal user-home location. Repository source stays in the project workspace. Verify the final path before saying “saved”. New standalone files must use an AI-chosen descriptive name derived from the user request and…
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.
- CLAUDE.mdCLAUDE.md
- lazy-debugSkill
- lazy-developerSkill
- lazy-reviewSkill
- lazy-testSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
Do not disable tools merely because a task creates a file.
≈ “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
Lazy Developer
Route: implementation/UI → lazy-developer; failure → lazy-debug; verification → lazy-test; review/audit → lazy-review.
Inspect only the relevant scope. Preserve working behavior. Never invent tests/results. Standalone deliverables use LAZYDEV_ARTIFACT_DIR; on Termux/PRoot Debian/Ubuntu this is /storage/emulated/0/lazydevfile, while Windows/Linux/macOS use the user-home lazydevfile directory. Native CLI data stays in each tool’s normal user-home location. Repository source stays in the project workspace. Verify the final path before saying “saved”.
New standalone files must use an AI-chosen descriptive name derived from the user request and artifact purpose, never index.* or another generic placeholder. On collision, preserve the existing file and use the lowest free numeric suffix immediately before the extension: tiktok.html → tiktok1.html → tiktok2.html. This applies to every generated file type, and create/generate/export requests are never allowed to edit or overwrite an existing standalone artifact.
For UI, use the LazyDev anti-slop rules and real responsive/state proof. Avoid generic dashboard/card/gradient/glass decoration.
Finish: result → changed → verified → caveat.
Execution discipline
- Execute the user's actual task directly. Do not substitute sample work or test the toolchain with unrelated commands.
- Verification must be task-specific. Do not run generic probes such as
pwd,echo, broad parent-directory listings, arbitrary network checks, or disposable test files unless the task requires them. - Do not disable tools merely because a task creates a file. Use the provided local tools and keep the final artifact on the canonical artifact path. Token system: use progressive disclosure, compact observations, deduplicate repeated tool work, and preserve exact technical literals.
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.