cricket-mode
AGENTS.md0 starsChanged 50 days agoMITView on GitHub
Cricket Mode is a lightweight, portable behavior layer for coding agents: a set of small, opt-in commands that help clamp scope, challenge assumptions, verify work, explain changes, and clean up without imposing a required workflow. Preserve the distinction between core behavior and platform adapters. Do not make a platform-specific behavior part of the core unless it is genuinely required for the command's meaning. Do not silently run every Cricket behavior on every task. When a user explicitly invokes a Cricket…
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.
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
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
- Distinctive: line 15
Do not make a platform-specific behavior part of the core unless it is genuinely required for the command's meaning.
- Distinctive: line 19
Do not silently run every Cricket behavior on every task.
- Very specific: line 21
When a user explicitly invokes a Cricket command, or clearly asks for that command's behavior, follow core/COMMANDS.md.
- Distinctive: line 27
Prefer deterministic evidence when software can answer the question: tests for test results, build tools for compilation, type checkers for …
- Distinctive: line 37
If a platform cannot support a behavior exactly, document the limitation instead of pretending parity exists.
≈ “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
Cricket Mode agent guidance
Cricket Mode is a lightweight, portable behavior layer for coding agents: a set of small, opt-in commands that help clamp scope, challenge assumptions, verify work, explain changes, and clean up without imposing a required workflow.
Project rule
Preserve the distinction between core behavior and platform adapters.
core/defines what a Cricket command means..agents/skills/contains the current working reference implementations.adapters/contains platform-specific glue.docs/PORTABLE-CRICKET.mdis the architecture and continuation plan.orchestration/is an optional policy beside the commands. Portable v1 does not require it.
Do not make a platform-specific behavior part of the core unless it is genuinely required for the command's meaning.
Commands are opt-in
Do not silently run every Cricket behavior on every task.
When a user explicitly invokes a Cricket command, or clearly asks for that command's behavior, follow core/COMMANDS.md.
The commands must remain independently useful.
Evidence rule
Prefer deterministic evidence when software can answer the question: tests for test results, build tools for compilation, type checkers for types, linters for lint, git diff for changes, and the real application path for user-visible behavior.
AI judgment may interpret evidence or handle uncertainty, but it should not replace a relevant deterministic check.
Portability rule
Adapters may translate command discovery, invocation syntax, instruction placement, tool wiring, and installation mechanics.
Adapters must not quietly redefine a command's purpose, scope boundary, evidence requirement, or result semantics.
If a platform cannot support a behavior exactly, document the limitation instead of pretending parity exists.
Keep Cricket small
Cricket Mode is not a mandatory agent framework.
orchestration/policy.py is the optional Phase 6 policy. It recommends a lane and, at most, existing Cricket commands. It does not select a model, run tests, or invoke those commands. Do not turn it into a router or an autonomous loop unless a later phase says so.
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.

