sonash-v0
AGENTS.md2 starsChanged 9 months agoApache-2.0View on GitHub
These instructions apply to Codex work in this repository. Claude Code remains a separate, supported development system. Preserve CLAUDE.md and .claude/; do not alter, migrate, or delete them unless the user explicitly requests it. Do not copy Claude transcripts, caches, secrets, or runtime state into Codex files. - Use Node.js 22 for the application and Functions work. The checked-in devcontainer is the reference environment. - This is a Next.js App Router application backed by Firebase and Cloud Functions. - Keep…
- Reads credentials
We split this AGENTS.md into the rules it gives an agent
11 rules. About 0 are common, 8 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Very specific: line 18
Use Node.js 22 for the application and Functions work.
- Very specific: line 22
Keep credentials in Codespaces secrets or ignored local files.
- Distinctive: line 24
Do not push branches or publish changes without explicit user approval.
- Distinctive: line 25
Keep dependency upgrades separate from Codex configuration work and use an isolated branch or worktree for major upgrades.
- Distinctive: line 33
Put shared Firestore queries in the repository service layer and use the existing domain types and runtime schemas.
- Distinctive: line 35
Follow the existing TypeScript, ESLint, Prettier, Tailwind, and Zod patterns.
- Very specific: line 36
Treat generated files and .claude/ runtime state as separate from product source changes.
- Distinctive: line 59
Use the Node 22 devcontainer for the full validation suite.
- Distinctive: line 71
Keep changes narrowly scoped and preserve unrelated user work.
- Distinctive: line 72
Run proportionate validation and report any skipped checks with the reason.
- Distinctive: line 73
Update durable project documentation when the supported workflow changes.
≈ “Common” is an estimate: at least 1 in 5 of the 36,368 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
SoNash Codex Instructions
<!-- prettier-ignore-start -->Document Version: 1.0 Last Updated: 2026-08-06 Status: ACTIVE
<!-- prettier-ignore-end -->Scope
These instructions apply to Codex work in this repository. Claude Code remains a
separate, supported development system. Preserve CLAUDE.md and .claude/; do
not alter, migrate, or delete them unless the user explicitly requests it. Do
not copy Claude transcripts, caches, secrets, or runtime state into Codex files.
Repository baseline
- Use Node.js 22 for the application and Functions work. The checked-in devcontainer is the reference environment.
- This is a Next.js App Router application backed by Firebase and Cloud Functions.
- Keep credentials in Codespaces secrets or ignored local files. Never commit
.env*.local, Firebase service-account files, or tokens. - Do not push branches or publish changes without explicit user approval.
- Keep dependency upgrades separate from Codex configuration work and use an isolated branch or worktree for major upgrades.
Application rules
- Route journal, daily-log, and inventory writes through the approved Cloud Functions surface; do not add direct client Firestore writes for these data paths.
- Put shared Firestore queries in the repository service layer and use the existing domain types and runtime schemas.
- Follow the existing TypeScript, ESLint, Prettier, Tailwind, and Zod patterns.
- Treat generated files and
.claude/runtime state as separate from product source changes. Do not clean or discard user changes to them without asking.
Common commands
From the repository root:
npm ci
npm --prefix functions ci
npm run type-check
npm run lint
npm test
npm run buildFor Functions-only validation:
npm --prefix functions run lint
npm --prefix functions run buildUse the Node 22 devcontainer for the full validation suite. If a host environment produces different results, report the environment and failure instead of rewriting the lockfile or weakening a check to make it pass.
For security-sensitive changes, run npm run security:secrets when gitleaks is
available. For hook or repository-automation changes, also run the relevant
npm run hooks:test, npm run hooks:health, or npm run test:gates command.
Change workflow
- Inspect the current worktree and relevant plan/context documents before editing.
- Keep changes narrowly scoped and preserve unrelated user work.
- Run proportionate validation and report any skipped checks with the reason.
- Update durable project documentation when the supported workflow changes.
- Review the final diff for secrets, generated-state churn, and accidental changes to Claude infrastructure.
Codex-specific boundary
Codex configuration and hooks belong under AGENTS.md and .codex/. Do not
silently port Claude hooks, agents, skills, MCP settings, or session mechanics.
<!-- BEGIN:nextjs-agent-rules -->This is NOT the Next.js you know
This version has breaking changes — APIs, conventions, and file structure may
all differ from your training data. Read the relevant guide in
node_modules/next/dist/docs/ (resolved from this file's directory; in
monorepos the next package may not be visible from the repo root) before
writing any code. Heed deprecation notices.
This block is written and re-added by next dev — verify at
node_modules/next/dist/server/lib/generate-agent-files.js. Removing it from a
diff only re-creates the uncommitted change; committing it with your work keeps
the tree clean.
<!-- END:nextjs-agent-rules -->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 28,693 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 repository81 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- llms.txtllms.txt
- find-skillsSkill
- add-debtSkill
- alertsSkill
- analyzeSkill
- artifacts-builderSkill
- audit-agent-qualitySkill
- audit-aggregatorSkill
- audit-ai-optimizationSkill
- audit-codeSkill
- audit-comprehensiveSkill
- audit-documentationSkill
- audit-engineering-productivitySkill
- audit-enhancementsSkill
- audit-healthSkill
- audit-performanceSkill
- audit-processSkill
- audit-refactoringSkill
- audit-securitySkill
- brainstormSkill
- checkpointSkill
- code-reviewerSkill
- comprehensive-ecosystem-auditSkill
- content-research-writerSkill
- convergence-loopSkill
- create-auditSkill
- data-effectiveness-auditSkill
- debt-runnerSkill
- decrypt-secretsSkill
- deep-planSkill
- deep-researchSkill
- developer-growth-analysisSkill
- doc-ecosystem-auditSkill
- doc-optimizerSkill
- docs-maintainSkill
- document-analysisSkill
- ecosystem-healthSkill
- Excel AnalysisSkill
- find-skillsSkill
- frontend-designSkill
- gh-fix-ciSkill
- github-healthSkill
- health-ecosystem-auditSkill
- hook-ecosystem-auditSkill
- market-research-reportsSkill
- mcp-builderSkill
- media-analysisSkill
- multi-ai-auditSkill
- pre-commit-fixerSkill
- pr-ecosystem-auditSkill
- pr-retroSkill
- pr-reviewSkill
- quick-fixSkill
- recallSkill
- repo-analysisSkill
- repo-synthesisSkill
- script-ecosystem-auditSkill
- session-beginSkill
- The first 60 shown.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

