agentleFS
thenavidmthenavidm/apple-photos-mcp-cli/AGENTS.md

apple-photos-mcp-cli

AGENTS.md1 starsChanged 50 days agoMITView on GitHub

See CLAUDE.md. One document, so the two cannot drift. src-ts/ is the MCP server and CLI; src/apple_photos_mcp/ is the engine. Python stays because osxphotos and photoscript are the only libraries that can read a Photos library and both are Python-only. TypeScript wraps them for the things Python cost us: npx with no toolchain, a CLI, and a tool array the HQ connector can import. Add a tool to Python and to src-ts/tools/index.ts together. A parity test compares the two lists…

The file

Files
AGENTS.md

Working on this repo

See CLAUDE.md. One document, so the two cannot drift.

The TypeScript surface

src-ts/ is the MCP server and CLI; src/apple_photos_mcp/ is the engine. Python stays because osxphotos and photoscript are the only libraries that can read a Photos library and both are Python-only. TypeScript wraps them for the things Python cost us: npx with no toolchain, a CLI, and a tool array the HQ connector can import.

Add a tool to Python and to src-ts/tools/index.ts together. A parity test compares the two lists and fails when they drift. That drift is what left the hosted connector exposing 11 of 13 tools, missing library_stats and look_at_photos, so a model estimated library totals from keyword samples and recommended photos it had never seen.

Compiled output goes to lib/, never dist/. Python packaging writes a dist/.gitignore containing *, which silently excludes every compiled file from the npm tarball. The package then installs with no code in it and the failure appears somewhere else entirely.

No os field in package.json. It reads as correct and breaks the HQ build: that route imports the tool array on a Linux builder and never executes a tool, so the package has to install anywhere. requireMac() refuses to run off macOS instead, which is the same call pyproject.toml makes with its platform markers.

npm run typecheck && npm test && npm run build
uv run pytest -q

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.