codex-toolkit
AGENTS.md2 starsChanged 11 days agoMITView on GitHub
This repository defines reusable Codex skills, agents, routing, installation, and evaluation behavior. Keep repository guidance small and durable; detailed task procedures belong in the relevant skill, orchestration document, script, or runbook. Use the narrowest relevant checks while iterating. Before an integrated toolkit change is complete, run the repository-provided validation for the affected skills/routing/installer and the normal package/CI checks. Do not weaken validators to make a new skill or workflow pass. A toolkit change is complete when routing remains unambiguous, installation/update…
1 other project uses a version of this.
Adopted by
Most recently taken up by
- cmdr-chara/chara-toolkitword for word, first seen Oct 13, 2026
Other versions1 project uses a version of this.
Word for word · 1
- cmdr-chara/chara-toolkit★ 2100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
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 31,843 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 repository34 other files it gives its agents.
- bin/AGENTS.mdAGENTS.md
- evaluations/AGENTS.mdAGENTS.md
- orchestration/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- skills/AGENTS.mdAGENTS.md
- skills/llms.txtllms.txt
- bug-finderSkill
- codebase-evolution-controllerSkill
- codebase-improvement-plannerSkill
- completion-proofSkill
- content-provenance-hygieneSkill
- cpp-engineeringSkill
- debugging-investigatorSkill
- delegate-with-mission-cardsSkill
- documentation-synchronizerSkill
- expo-react-native-builderSkill
- flutter-production-builderSkill
- go-engineeringSkill
- kotlin-engineeringSkill
- mobile-architecture-directorSkill
- multi-agent-work-coordinatorSkill
- optimize-codebase-performanceSkill
- product-design-directorSkill
- production-web-builderSkill
- python-engineeringSkill
- repository-intelligenceSkill
- review-and-refactor-codeSkill
- rust-engineeringSkill
- screenshot-to-interfaceSkill
- security-reviewSkill
- swift-engineeringSkill
- toolchain-preflightSkill
- typescript-quality-enforcerSkill
- verification-and-releaseSkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 21
Prefer editing an existing skill or router when the concept already has an owner.
- Distinctive: line 22
Keep skill descriptions precise about when the skill applies.
- Distinctive: line 23
Do not encode model weaknesses as permanent ceremony when the behavior can be expressed as a durable constraint, permission boundary, …
- Very specific: line 28
Use the narrowest relevant checks while iterating.
≈ “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
Chara's Toolkit agent instructions
Purpose
This repository defines reusable Codex skills, agents, routing, installation, and evaluation behavior. Keep repository guidance small and durable; detailed task procedures belong in the relevant skill, orchestration document, script, or runbook.
Core invariants
- Route to the smallest specialist/workflow that owns the task. Do not turn optional workflow edges into mandatory chains.
- Repository-local
AGENTS.mdinstructions remain authoritative over generic toolkit guidance. - Specialist safety, approval, migration, and stop conditions cannot be bypassed by orchestration or completion helpers.
- Treat destructive filesystem, Git, database, and shell operations as a separate authorization boundary on Windows, macOS, and Linux—even in full-access mode. Check literal resolved targets and ownership before acting; never delete drive/filesystem roots, home directories, repository roots, backups, or unrelated work. Do not retry a blocked action using another shell, language, or tool.
completion-proofowns completion discipline;verification-and-releaseowns integrated ship/no-ship judgment. Keep those responsibilities distinct.- Managed global
AGENTS.mdinstallation must preserve all user content outside the toolkit markers and fail closed on malformed/duplicate managed blocks. - Auto-update behavior follows published releases, not unreleased
maincommits. - Managed enterprise installs require externally signed package approval, disable scheduled updates, and preserve host authorization boundaries.
- Release publication requires an explicitly approved, CI-verified candidate; no automatic patch publication.
Changes
- Prefer editing an existing skill or router when the concept already has an owner; avoid overlapping skills with near-identical triggers.
- Keep skill descriptions precise about when the skill applies. Put long procedures and references behind progressive disclosure instead of front-loading them into descriptions or global routing.
- Do not encode model weaknesses as permanent ceremony when the behavior can be expressed as a durable constraint, permission boundary, trigger, or completion condition.
- Preserve installer idempotence, custom
CODEX_HOME, and cross-platform behavior when changing packaging or setup code.
Verification
Use the narrowest relevant checks while iterating. Before an integrated toolkit change is complete, run the repository-provided validation for the affected skills/routing/installer and the normal package/CI checks. Do not weaken validators to make a new skill or workflow pass.
Completion
A toolkit change is complete when routing remains unambiguous, installation/update behavior is preserved, affected validators/evaluations pass, public docs reflect user-visible changes, and any compatibility or rollout risk is stated explicitly.
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.