ontology-atlas
CLAUDE.md156 starsChanged 5 days agoMITView on GitHub
AGENTS.md is canonical. This wrapper adds only what differs for Claude Code: loading and hook ownership. Anything true for both harnesses belongs in AGENTS.md. .claude/rules/ contains always-loaded forbidden, git, and local-first rules. Path-loaded rules are design, design-gates, architecture, testing, surfaces, and documentation; their frontmatter paths and rules-path-scope.contract.test.ts keep the conditions live. Making a rule resident updates that contract, this sentence, and the commit's reason together. Opus builds (implementer) and debugs (investigator); Fable plans, specs, reviews and gives design verdicts (docs/engineering/agent-effort.md).…
In this repository28 other files it gives its agents.
- AGENTS.mdAGENTS.md
- app/AGENTS.mdAGENTS.md
- cli/AGENTS.mdAGENTS.md
- mcp/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- src/AGENTS.mdAGENTS.md
- src-tauri/AGENTS.mdAGENTS.md
- tests/AGENTS.mdAGENTS.md
- delegateSkill
- gate-probeSkill
- mergeSkill
- ontology-bootstrapSkill
- ontology-extractSkill
- ontology-field-trialSkill
- ontology-syncSkill
- security-auditSkill
- ui-buildSkill
- ui-proofSkill
- delegateSkill
- gate-probeSkill
- mergeSkill
- ontology-bootstrapSkill
- ontology-extractSkill
- ontology-field-trialSkill
- ontology-syncSkill
- security-auditSkill
- ui-buildSkill
- ui-proofSkill
The file
Files
CLAUDE.md
AGENTS.md is canonical. This wrapper adds only what differs for
Claude Code: loading and hook ownership. Anything true for both harnesses
belongs in AGENTS.md.
@AGENTS.md
Claude Code loading
.claude/rules/ contains always-loaded forbidden, git, and
local-first rules. Path-loaded rules are design, design-gates,
architecture, testing, surfaces, and documentation; their frontmatter
paths and rules-path-scope.contract.test.ts keep the
conditions live. Making a rule resident updates that contract, this sentence,
and the commit's reason together.
Opus builds (implementer) and debugs (investigator); Fable plans, specs,
reviews and gives design verdicts (docs/engineering/agent-effort.md).
Claude discovers skills under .claude/skills/.
Codex discovers .agents/skills/; it does not auto-load CLAUDE.md.
Hooks
.claude/settings.json owns Claude permissions and hooks. Every hook is
mirrored in .codex/hooks/ as an adaptation (a Codex edit is an apply_patch
envelope), not a copy; block-secret-read.sh is Codex-only because Claude uses
permissions.deny, and session-length.sh is Claude-only because it reads
Claude's transcript. Hooks block irreversible commands, report lint and language
findings at edit time, and inject the vault census at session start; each
header says why. pnpm test:claude:hooks guards the wiring.
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.

