agentleFS
suxrobGMsuxrobGM/jobpilot/CLAUDE.md

jobpilot

CLAUDE.md113 starsChanged 9 days agoMITView on GitHub

Multi-user AI job-application app. The web UI and API are cloud-hosted and shared. The API owns all state. Each user runs the agent locally: Claude Code or Codex in a .NET PTY host, plus Playwright. Dev ports: web :4100, API :4101, PTY host :4102. Auth: the web signs in with an opaque session (httpOnly session cookie, a row in auth_tokens). It fetches the user's terminal token via POST /api/auth/tokens/terminal and hands it to the PTY host, which sets JOBPILOT_API_TOKEN. Skills…

The file

We split this CLAUDE.md into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are 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.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Never run biome check --write --unsafe.

    line 43
  2. Very specific:

    Run bun run verify before committing, plus dotnet test when apps/terminal/ or tests/ changed.

    line 46
  3. Distinctive:

    No IIFEs. No fallback or compat shims.

    line 52
  4. Distinctive:

    No nested ternaries. Everyday names over jargon.

    line 53
  5. Distinctive:

    Add a test only when it guards behavior that could realistically break.

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

JobPilot

Multi-user AI job-application app. The web UI and API are cloud-hosted and shared. The API owns all state. Each user runs the agent locally: Claude Code or Codex in a .NET PTY host, plus Playwright. Dev ports: web :4100, API :4101, PTY host :4102.

Auth: the web signs in with an opaque session (httpOnly session cookie, a row in auth_tokens). It fetches the user's terminal token via POST /api/auth/tokens/terminal and hands it to the PTY host, which sets JOBPILOT_API_TOKEN. Skills call the API through plugin/bin/jobpilot-api, which sends it as Authorization: Bearer. Every bearer call (host and skills) also sends x-jobpilot-version; the API answers 426 below MIN_AGENT_VERSION (apps/api/src/common/middleware/agent-version.ts). Raise it when a release removes a route.

Layout

PathWhat it is
apps/web/Next.js 16 + MUI 9 on Bun. Talks to the API over HTTP only. src/proxy.ts is auth middleware, not a data proxy.
apps/api/Elysia + Prisma 7 on Bun. Owns all persistence. Exports type App for Eden. Swagger at :4101/swagger.
apps/terminal/.NET 10 PTY host. Tests in tests/JobPilot.Terminal.Tests/, solution JobPilot.slnx.
packages/contracts/@jobpilot/contracts: shared Zod schemas.
packages/api-client/@jobpilot/api-client: Eden client.
plugin/One skill tree for Claude and Codex. Edit skills here directly.
docs/, deploy/User docs. Production stack.

Commands (bun run <name>)

CommandWhat it does
devStart all three apps. dev:api / dev:web / dev:terminal for one.
db:tunnelSSH tunnel to remote PostgreSQL on localhost:5433.
db:setupPrisma generate + migrate + seed.
testAPI tests + contracts tests.
build:web / build:terminalProduction builds. The API runs from source.
checkBiome format + lint + import sort. Writes.
cibiome ci --error-on-warnings.
knipDead files, exports, and dependencies.
verifyThe CI gate: ci, knip, API and web typechecks, test.

A local PostgreSQL and the tunnel both use port 5433. Check docker ps or lsof -i :5433 before assuming which one DATABASE_URL hits.

Never run biome check --write --unsafe. It rewrites cookie[KEY]!.set(…) to ?.set(…) and drops auth cookie writes.

Run bun run verify before committing, plus dotnet test when apps/terminal/ or tests/ changed. Commit with the commit skill.

Code style

  • Comment only a non-obvious why (constraint, trap, rejected alternative). One line, four max.
  • No IIFEs. No fallback or compat shims; write a data migration.
  • No nested ternaries. Everyday names over jargon.
  • Export only what another file imports. bun run knip fails otherwise.
  • A barrel is allowed only for a directory's public API with several outside importers, using named re-exports. export * only for a package.json exports target.
  • Add a test only when it guards behavior that could realistically break. No redundant, duplicate, or meaningless tests (restating the implementation, re-testing what another test covers).
  • Split a test file past a few hundred lines by domain. Shared fixtures go in fakes.ts or builders.ts.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.