agentleFS
badgidsbadgids/comfyui-development-skills/skills/comfyui-core-runtime/SKILL.md

comfyui-core-runtime

Skill4 starsChanged 44 days agoMITView on GitHub

Use when changing or diagnosing ComfyUI core execution, server behavior, backend APIs, model loading, caching, scheduling, node execution internals, filesystem handling, or other code owned by the ComfyUI runtime rather than a third-party custom-node pack.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 8 are distinctive, and 1 is 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.

  • 1

    Very specific

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

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

    If the comfyui-development router skill is installed, follow its shared authority/freshness policy.

    line 9
  2. Distinctive:

    Search for the behavior and its tests rather than assuming old module paths.

    line 22
  3. Distinctive:

    Check for feature flags, version guards, compatibility shims, or deprecations.

    line 24
  4. Distinctive:

    Run the narrowest relevant tests, then broader tests if feasible.

    line 26
  5. Distinctive:

    use an RFC as proof that code has shipped.

    line 34
  6. Distinctive:

    update the user's checkout to upstream without permission.

    line 36
  7. Distinctive:

    Before calling the change complete:

    line 49
  8. Distinctive:

    verify public route, node, or workflow behavior when the change affects one.

    line 54
  9. Distinctive:

    update public documentation when the supported contract changed.

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill 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
SKILL.md

ComfyUI Core Runtime

If the comfyui-development router skill is installed, follow its shared authority/freshness policy. This specialist remains usable on its own.

Primary sources

  • target/version-matched Comfy-Org/ComfyUI;
  • matching/current Comfy-Org/docs;
  • tests in the core repository;
  • RFCs only for proposal/history context.

Procedure

  1. Establish the compatibility target and exact checkout/commit.
  2. Reproduce or locate the owning subsystem before editing.
  3. Search for the behavior and its tests rather than assuming old module paths.
  4. Trace call sites across the execution path.
  5. Check for feature flags, version guards, compatibility shims, or deprecations.
  6. Make the smallest behaviorally complete change.
  7. Run the narrowest relevant tests, then broader tests if feasible.
  8. For server/API changes, verify the actual route/payload behavior on a running target when practical.
  9. For node-facing changes, inspect resulting live node definitions or representative workflows.

Do not

  • hardcode an internal path merely because it existed in a previous release;
  • infer an API contract from an internal helper alone;
  • use an RFC as proof that code has shipped;
  • refactor unrelated code as part of a bug fix;
  • update the user's checkout to upstream without permission.

Evidence for completion

At least one of these should exist before claiming the change works:

  • relevant automated tests pass;
  • a minimal runtime reproduction passes;
  • a server/API probe confirms the behavior;
  • a representative workflow exercises the changed path.

Acceptance gate

Before calling the change complete:

  • record or identify the target ComfyUI commit/version;
  • show that the change addresses the owning execution path;
  • pass a focused test or minimal runtime reproduction;
  • verify public route, node, or workflow behavior when the change affects one;
  • update public documentation when the supported contract changed.

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.