agentleFS
badgidsbadgids/Story-Film-Skills/skills/generation-workflow-setup/SKILL.md

generation-workflow-setup

Skill4 starsChanged 3 months agoApache-2.0View on GitHub

Discover complete ComfyUI workflows only from Story-Film's extension-local comfyui_workflows library, then let the user select each required task from an ordinary numbered list.

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. About 0 are common, 7 are distinctive, and 9 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 9

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill before ComfyUI generation or whenever the user wants to change the workflow used for a generation task.

    line 11
  2. Distinctive:

    When the selected playbook will use ComfyUI, workflow selection happens before story/canon creation.

    line 19
  3. Distinctive:

    If a required category already has a durable selection and the user did not ask to change it, reuse that exact selection.

    line 21
  4. Very specific:

    Only missing categories or an explicit user-requested workflow change open an interactive numbered catalog.

    line 23
  5. Very specific:

    Check 00_project/workflow_preferences.json first.

    line 27
  6. Very specific:

    Add --query <text> when the task needs a narrower subset.

    line 30
  7. Very specific:

    Do not use ask_user_question or another TUI picker for workflow selection.

    line 32
  8. Distinctive:

    Do not truncate the list. Story-Film imposes no maximum workflow count.

    line 33
  9. Distinctive:

    Ask the user to reply with the number of the workflow to use.

    line 35
  10. Very specific:

    After the user replies, record that exact choice with workflow_catalog.py choose.

    line 36
  11. Distinctive:

    If the workflow names unavailable models or missing nodes, report the exact blocker.

    line 39
  12. Distinctive:

    Do not prefer a source merely because it appears first.

    line 51
  13. Very specific:

    Do not run a second adapter/checkpoint/VAE/text-encoder/LoRA questionnaire after workflow selection.

    line 55
  14. Very specific:

    When checking whether resource names stored in the selected workflow are available, use the running ComfyUI server registry.

    line 63
  15. Very specific:

    Use scripts/model_inventory.py scan only for low-level diagnostics and compatibility reporting.

    line 65
  16. Very specific:

    Do not add a generate-new choice to every normal workflow catalog.

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Generation Workflow Setup

Use this skill before ComfyUI generation or whenever the user wants to change the workflow used for a generation task.

Read

Read ../../references/WORKFLOW_SELECTION.md.

Playbook preflight and reuse

When the selected playbook will use ComfyUI, workflow selection happens before story/canon creation. Run ../../scripts/workflow_preflight.py set for the playbook/profile or explicit task categories, then resolve every missing category before creative production begins.

If a required category already has a durable selection and the user did not ask to change it, reuse that exact selection. Materialize or validate it as needed, but do not show the numbered list again. Do not ask again later.

Only missing categories or an explicit user-requested workflow change open an interactive numbered catalog. A later missing model/node/input is a blocker, not permission to reopen workflow selection.

Required behavior

  1. Check 00_project/workflow_preferences.json first. If the task already has a durable selection and no change was requested, reuse it without another user question.
  2. Determine the generation task category that the current preflight or production step needs.
  3. For a missing category, run ../../scripts/workflow_catalog.py catalog <project-root> --category <task>.
  4. Add --query <text> when the task needs a narrower subset.
  5. Show the command's complete ordinary numbered list to the user.
  6. Do not use ask_user_question or another TUI picker for workflow selection.
  7. Do not truncate the list. Story-Film imposes no maximum workflow count. Show every built-in and package-custom workflow in the extension library regardless of how many choices exist. There is no four-choice limit for workflow selection. There is no higher Story-Film workflow-count limit.
  8. Ask the user to reply with the number of the workflow to use.
  9. After the user replies, record that exact choice with workflow_catalog.py choose.
  10. Materialize the selected extension source with workflow_catalog.py materialize.
  11. Inspect and live-validate the selected workflow before execution.
  12. If the workflow names unavailable models or missing nodes, report the exact blocker. Do not silently replace the workflow, model, VAE, encoder, LoRA, or other resource.
  13. Continue to the next required task category only after the current workflow choice is recorded.

Source types

The catalog can include only:

  • bundled workflows under ../../comfyui_workflows/<task>/<model>/;
  • package custom workflows under ../../comfyui_workflows/custom/<task>/<model>/.

Project workflow/default folders, ComfyUI userdata, arbitrary external paths, and ComfyUI template catalogs are not discovery sources.

Do not prefer a source merely because it appears first. The user chooses from the numbered list.

Workflow owns the model stack

Do not run a second adapter/checkpoint/VAE/text-encoder/LoRA questionnaire after workflow selection.

The selected workflow owns the concrete model/resource values stored in its graph. Prompt adapters may still translate model-neutral production intent when the selected workflow requires model-specific prompt grammar.

model_preferences.json is legacy compatibility data. It does not override a selected workflow.

Live model registry validation

When checking whether resource names stored in the selected workflow are available, use the running ComfyUI server registry. Use /models and /models/{folder} through the bundled discovery helpers. These registry results include external model roots configured by extra_model_paths.yaml.

Use scripts/model_inventory.py scan only for low-level diagnostics and compatibility reporting. Do not run find / or another filesystem sweep to rediscover model files. Do not replace it with direct curl, wget, or one-off Python parsers. The running ComfyUI registry is runtime truth for model availability.

Custom workflow location

To add any custom, edited, exported, or newly authored workflow to Story-Film, copy its JSON file under:

comfyui_workflows/custom/<task>/<model>/

Refresh the catalog after copying it. Story-Film never registers or scans an arbitrary external workflow path.

Explicit workflow creation

Do not add a generate-new choice to every normal workflow catalog. However, if the user explicitly asks Story-Film to create, build, author, or design a new ComfyUI workflow, route to comfyui-workflow and use its bounded live-schema authoring path.

The new candidate must be built from real live node/model schemas, validated before promotion, and must not invent node classes or silently install dependencies. If the new workflow should become reusable/selectable by Story-Film, save or copy the validated JSON into comfyui_workflows/custom/<task>/<model>/, refresh the catalog, and record that real workflow as the durable selection when appropriate.

Done

Done when each generation task required by the current production scope has a durable selection in 00_project/workflow_preferences.json, any active workflow preflight is complete, and the selected workflow has either passed its required validation or has a concrete unresolved blocker.

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.