Skip to content

Since the last run

Overall +30 points

comfyanonymous/ComfyUI (opens in a new tab)

A heavyweight AI tool with strong community presence, but setup and navigation have real friction for newcomers.

Python★ 117,0984,130 open issuesGNU General Public License v3.0
Overall score: 72 out of 100
Holding

last scored 2026-06-15

Readiness lenses

The same repo, two ways: how ready it is for an AI coding agent, and for a human newcomer. Switch lenses to see each one's score and checks.

How ready the repo is for an AI coding agent to read, build, and operate it.

16 of 32 checks passed across 6 categories.

  • Machine Orientation

    4 checks passed3 problems found

    • mediumAgent instructions file

      AGENTS.md / CLAUDE.md is the direct, machine-read channel for entry points and conventions.

      Fix: Add an AGENTS.md naming the entry points, the build/test commands, and the house conventions.

    • lowllms.txt present

      An llms.txt is the emerging machine-orientation standard — a curated map for LLMs.

      Fix: Add an llms.txt pointing at the key docs and entry points.

    • lowArchitecture doc

      An ARCHITECTURE doc hands an agent the module map instead of making it infer one.

      Fix: Add an ARCHITECTURE.md describing the major components and how they fit.

    • README present — An agent reconstructs what the repo is and how to work in it from the README first.
    • Repository description set — The one-line description is the fastest orientation signal an agent reads.
    • Topics set — GitHub topics situate the repo's domain and stack for an agent at a glance.
    • Contribution conventions are written down — A contributor guide is where an agent learns the branch, commit, and review conventions it must follow.
  • Reproducible Setup

    0 checks passed5 problems found

    • highLockfile committed

      Without a lockfile two clones can resolve different dependency trees — the build isn't reproducible.

      Fix: Commit your lockfile (package-lock.json / pnpm-lock.yaml / Cargo.lock / …).

    • mediumRuntime / toolchain pinned

      An unpinned runtime makes an agent guess the version a clean environment needs.

      Fix: Pin the runtime with .nvmrc / .tool-versions / an engines field / packageManager.

    • mediumContainer or dev-container

      A Dockerfile / compose / devcontainer removes 'works on my machine' from setup.

      Fix: Add a Dockerfile, a compose file, or a .devcontainer so the environment is declared, not assumed.

    • lowEnv template present

      A .env.example tells an agent which configuration keys it must supply to run.

      Fix: Ship a .env.example listing every required key with placeholder values.

    • lowDev container

      A .devcontainer gives an agent a one-command, fully-declared environment.

      Fix: Add a .devcontainer so the toolchain is declared, not assumed.

  • Code Navigability

    1 check passed0 problems found

    • Conventional source layout — A recognised src/ · lib/ · app/ layout lets an agent localise code without spelunking.
  • Verification Surface

    2 checks passed5 problems found

    • mediumTest command defined

      A `test` script is the documented, discoverable way to run the suite.

      Fix: Add a `test` script (package.json / Makefile / Taskfile) that runs the suite.

    • lowLint command defined

      A linter an agent can run catches a class of mistakes before they ship.

      Fix: Add a `lint` script and a linter config.

    • lowTask runner present

      A Makefile / Taskfile / justfile gives an agent a single index of the project's commands.

      Fix: Add a Makefile / Taskfile / justfile collecting the common commands.

    • lowTypecheck command

      A typecheck script lets an agent catch type errors before they ship.

      Fix: Add a `typecheck` script (e.g. `tsc --noEmit`).

    • lowBuild command

      A documented build command is the discoverable way for an agent to produce artifacts.

      Fix: Add a `build` script that produces the distributable.

    • Tests present — Runnable tests are the ground truth an agent checks its own edits against.
    • CI configured — Readable CI is the feedback loop an agent reads to know what 'green' means here.
  • Agent Safety

    6 checks passed2 problems found

    • medium.gitignore covers env files

      An ignore rule for .env is what keeps the next careless commit from leaking secrets.

      Fix: Add `.env` (and `.env.*`) to .gitignore.

    • lowAutomated dependency updates

      Dependabot / Renovate keeps dependencies patched without a human in the loop.

      Fix: Enable Dependabot or Renovate.

    • No committed .env — A tracked real .env hands secrets to anything — including an agent — that reads the repo.
    • No exposed secrets — Secret-shaped literals in tracked files are a live credential leak.
    • CODEOWNERS present — CODEOWNERS marks the 'don't touch without review' zones an agent should respect.
    • .gitignore present — A .gitignore is the first guard against an agent committing secrets or build junk.
    • Security policy — A SECURITY.md is where an autonomous security agent reads this repo's rules before acting on it.
    • License makes the code reusable — Without a license an agent cannot safely copy, redistribute, or vendor any of this code.
  • Context Economy

    3 checks passed1 problem found

    • lowNo oversized source file

      A single file in the hundreds of KB blows past an agent's effective context window.

      Fix: Split the largest files along their natural seams.

    • Dependencies not vendored in — A committed node_modules / vendor tree drowns the real signal an agent needs.
    • Build output not committed — Generated dist/ · build/ output is noise an agent has to read past.
    • Generated / vendored files marked — A .gitattributes linguist marker tells tools (and agents) what to skip.

Where the grip slips

  1. highPath To Contribute+13 pts if fixed

    CONTRIBUTING.md lacks practical dev setup guide

    CONTRIBUTING.md directs questions to Discord, explains bug reporting, and links to an external wiki but does NOT include: how to install the project in editable/dev mode, how to run tests locally, how to start the app for development, what the project's code standards are (beyond linting config in pyproject.toml), or a step-by-step first-contribution walkthrough.

    Fix: Expand CONTRIBUTING.md with a 'Development Setup' section: (1) clone and install with `pip install -e .`, (2) run unit tests with `pytest`, (3) launch the dev server with `python main.py`, (4) where to find code style guidelines (refer to ruff/pylint config already in place). Add a 'Making Your First Contribution' subsection with an example (e.g., fixing a typo, writing a small node, submitting a test).

  2. highGetting Running+12 pts if fixed

    No embedded quickstart in README; installation path is indirect

    The README links to external install methods (Desktop app, portable package, manual) without providing a direct 'git clone + one-line command' entry point. A newcomer must click links or navigate external docs to start, creating friction and reducing discoverability of the local development path.

    Fix: Add a 'Quick Start' section in the README with copy-pasteable commands: e.g., 'git clone <url>, cd ComfyUI, python -m pip install -e . && python main.py' (or whatever the correct entry point is). Include system requirements (Python 3.10+, CUDA/GPU optional) and expected output.

  3. highFinding Your Way+11 pts if fixed

    No architecture overview or module-level docs in the codebase

    The README lists features but does not explain how the system is organized: where the core node engine lives, how the API server and frontend interact, what the asset database does, where custom node loading happens, or how the workflow system works. A newcomer cloning the repo sees comfy/, app/, api_server/, blueprints/, etc., but has no mental model of the system.

    Fix: Create an ARCHITECTURE.md at the root with a 3–5 paragraph overview: (1) Core node execution engine (comfy/), (2) API server and routes, (3) Frontend integration, (4) Asset/workflow database, (5) Custom node loading and app plugins. Include a dependency graph diagram and pointers to key entry points (e.g., main.py, key classes).

  4. highGetting Running+10 pts if fixed

    Unpinned dependencies in requirements.txt inhibit reproducible builds

    Key dependencies like torch, torchvision, torchaudio, numpy, transformers, safetensors, aiohttp, and many others lack version pinning (or use only >= constraints), making it impossible to guarantee the same environment across machines or over time. This is especially critical for a GPU-heavy project where version mismatches can cause subtle failures.

    Fix: Generate and commit a requirements-lock.txt (or poetry.lock / pip-compile output) with exact pinned versions of all dependencies and transitive dependencies. Update the Getting Started guide to recommend installing from the lock file. Include a script or CI step to keep it fresh.

  5. mediumPath To Contribute+8 pts if fixed

    No issue/PR templates for core code contributions; API-node-only template

    The .github/PULL_REQUEST_TEMPLATE/ contains api-node.md but no general template for core ComfyUI changes. This leaves contributors guessing about what to document, test, or explain in their PR description. The lack of beginner-friendly issue labels or 'good first issue' tracking makes it hard to find entry points.

    Fix: Create .github/PULL_REQUEST_TEMPLATE.md with sections for: (1) what changed and why, (2) related issue number, (3) testing checklist (unit tests, manual test steps), (4) breaking changes or migration notes. In the issues list, add 'good-first-issue' and 'help-wanted' labels to tractable problems suitable for newcomers.

  6. mediumGetting Running+7 pts if fixed

    Manual install section in README is truncated; full steps are not visible

    The README text ends abruptly with 'R' under the keyboard shortcuts, suggesting the manual installation instructions are cut off or missing from the snapshot. Without access to the full guide, a newcomer following the README directly cannot complete setup without hunting for docs elsewhere.

    Fix: Ensure the full Manual Install section is present and complete in the README. Include OS-specific commands for Windows, Linux, and macOS, plus GPU-specific guidance (NVIDIA, AMD, Intel, Apple Silicon). If the section is too long, move it to INSTALL.md and link from the README.

  7. lowOpen Door+4 pts if fixed

    No code of conduct or behavioral guidelines for the community

    While Discord and Matrix channels are active and linked, there is no CODE_OF_CONDUCT.md or GOVERNANCE.md in the repo to set expectations for respectful, inclusive participation. This is a light risk for a large open-source project and can lead to friction or unclear resolution when conflicts arise.

    Fix: Add a CODE_OF_CONDUCT.md file (e.g., using the Contributor Covenant) and link it prominently in the README and CONTRIBUTING.md. Keep it concise and include a clear path to report violations (e.g., email or Discord moderation team).

Show your score

Latch badge
[![Latch score](https://latch.fyi/api/badge/dd2c740a-f887-4ccf-9014-ee897f7bce01)](https://latch.fyi/comfyanonymous/comfyui)

Share this analysis

https://latch.fyi/comfyanonymous/comfyui

Fixed some issues? Re-analyze to measure your progress →

Copy the whole report

Grab the entire verdict in one go — a Markdown block for a README or an issue, plain text, or a prompt you can paste straight into an LLM.

Was this verdict fair?

One tap helps — or tell us more.