Agents
AGENTS.md1 starsChanged 58 days agoApache-2.0View on GitHub
Claude Code never reads anything under this directory. That is documented behavior, not a bug. Nothing here loads in a Claude Code session: not hooks.json, not agents/, not skills/, not rules/, and not this file. The tree exists for other agent harnesses. ../.claude/ is the authoritative configuration for Claude Code, and when the two disagree the .claude/ copy is the one that runs. - Editing a file here changes nothing about a Claude Code session. If you meant to change…
In this repository39 other files it gives its agents.
- agent-core/AGENTS.mdAGENTS.md
- AGENTS.mdAGENTS.md
- behavioral-regression/AGENTS.mdAGENTS.md
- claude-foundation/AGENTS.mdAGENTS.md
- config/AGENTS.mdAGENTS.md
- corpora/AGENTS.mdAGENTS.md
- demo/AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- examples/AGENTS.mdAGENTS.md
- experiments/AGENTS.mdAGENTS.md
- flow-corpus/AGENTS.mdAGENTS.md
- flow-protocol/AGENTS.mdAGENTS.md
- openspec/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- skills/AGENTS.mdAGENTS.md
- tests/AGENTS.mdAGENTS.md
- claude-foundation/CLAUDE.mdCLAUDE.md
- merge-gate-auditorSkill
- update-executive-reportSkill
- update-executive-reportSkill
- architecture-drift-guardSkill
- commonSkill
- corpus-guardianSkill
- dataset-lintSkill
- deploySkill
- e2e-matrix-sentinelSkill
- eval-corpus-forgeSkill
- hierarchical-recursive-brainstormSkill
- model-benchSkill
- openai-judgeSkill
- openspec-implementation-reviewSkill
- openspec-peer-reviewSkill
- openspec-quality-planSkill
- pre-pr-gateSkill
- project-setupSkill
- quality-gateSkill
- refactoring-decomposerSkill
- repo-invariant-reviewSkill
- test-completeness-guardSkill
The file
Files
AGENTS.md — .agents
Claude Code never reads anything under this directory. That is documented behavior, not a bug.
Nothing here loads in a Claude Code session: not hooks.json, not agents/, not
skills/, not rules/, and not this file. The tree exists for other agent harnesses.
../.claude/ is the authoritative configuration for Claude Code, and when the two disagree
the .claude/ copy is the one that runs.
Map
| Path | Role |
hooks.json | A different schema: a top-level object keyed by named guard, each holding its own event blocks |
agents/merge-gate-auditor.md | Byte-identical to the .claude/agents/ copy (verified with diff) |
skills/update-executive-report/ | Identical to the .claude/skills/ copy (verified with diff -r) |
skills/merge-gate-auditor/SKILL.md | Exists only here; no .claude/skills/ counterpart |
rules/code_standards.md | Exists only here |
Diagram
flowchart LR
accTitle: which harness reads which directory
accDescr: Claude Code reads only the .claude directory, while other harnesses read .agents; the two trees hold near-duplicate files and only the .claude copies ever take effect in a Claude Code session.
CC["Claude Code"]
OTHER["another agent harness"]
CLD[".claude/<br/>settings.json, hooks, agents"]
subgraph AGD[".agents/ (you are here)"]
HJ["hooks.json<br/>guard-keyed schema"]
AA["agents/merge-gate-auditor.md<br/>byte-identical copy"]
SK["skills/<br/>one duplicate, one unique"]
RU["rules/code_standards.md"]
end
CC --> CLD
CC -. never reads .-> AGD
OTHER --> HJ
OTHER --> AA
OTHER --> SK
OTHER --> RU
classDef here fill:#e8f0fe,stroke:#1a73e8,stroke-width:2px
class AGD hereRules that bite here
- Editing a file here changes nothing about a Claude Code session. If you meant to change
Claude Code's behaviour, the file you want is under
../.claude/. This is the single most expensive mistake available in this directory, because the edit looks successful. - The duplicates are real duplicates and there is no drift guard.
agents/merge-gate-auditor.mdis byte-identical to its.claude/twin andskills/update-executive-report/matches its.claude/twin exactly, but nothing in CI keeps them that way. Change one and you must change the other by hand, or the two harnesses quietly diverge. hooks.jsonis notsettings.jsonin a different place. It is keyed by guard name rather than by event, and its matchers name tool identifiers that Claude Code does not use. Copying wiring between the two files without translating it produces a hook that never fires, in either harness.- Some content here has no
.claude/counterpart.rules/code_standards.mdandskills/merge-gate-auditor/SKILL.mdexist only in this tree, so a Claude Code session has never seen them. Promote anything that should apply into../.claude/or the repo root.
Verify
diff -r .agents/skills/update-executive-report .claude/skills/update-executive-reportSubagents
| Task in this directory | Agent | Why |
Find every reference to a .agents/ path elsewhere in the repo | explorer | Read-only; stale prose pointing here is the usual source of the confusion |
| Diff the duplicated pairs and report which drifted | test-runner | Has Bash; diff is the only check there is, and nothing in CI runs it |
Review a change that promotes a rule into .claude/ | narrow-critic | Content written for another harness can assume tools this one does not have |
See also
| Doc | Read it when |
../.claude/README.md | You want to change what a Claude Code session actually does |
../AGENTS.md | You need the repo-wide orientation that every session does load |
../docs/plans/agents-md-directory-docs/PLAN.md | You want the evidence for the claim that this directory is never read |
agents/merge-gate-auditor.md | You are editing the merge-gate auditor role and must update both copies |
../docs/runbooks/merge-gate-audit-triage.md | You are actually running a merge-gate audit rather than editing its description |
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.

