JMCAI Comfypet
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md to install @user_8dd3bc37/comfypet-jmcai-skill.
About this skill
Problem
JMCAI Comfypet is aimed at connecting multimodal workflows to a local Agent without hand-writing workflow, target, or parameter structures. It addresses the practical problem of choosing a suitable workflow from summary, tags, input_modalities, output_modalities, and example prompts, then filling only the parameters exposed by the alias schema and keeping local or remote bridge asset upload, run polling, and error reporting in one stable flow.
How It Works
The skill depends on the JMCAI Comfypet Desktop Application. Its execution usually follows these steps:
- Select a workflow: inspect metadata and match the required input/output modalities.
- Fill parameters: use only alias names returned by registry --agent; when a required: true value is missing, ask the user or infer it from context.
- Submit the run: after receiving run_id, make a single status query instead of tight polling.
- Collect results: on success, prefer local file paths already downloaded to the current machine; on failure, preserve the original error message and do not fake success.
Boundaries and Notes
By default it communicates with 127.0.0.1. When bridge_url points to another desktop app on the LAN, the skill automatically uploads common assets such as image, mask, video, audio, and file as upload:. Remote uploads are limited to a whitelist covering media, txt/csv/pdf/srt/vtt/ass, and similar formats. Do not modify schema, workflow, or target configuration. If doctor fails, prompt the user to start or upgrade the desktop app; if status returns queued or running, continue in a later independent call.
Use Cases
- Select a Comfypet workflow from prompts, modalities, and tags, then submit it using only alias-schema parameters.
- After generating image, video, or audio output on the desktop app, poll by run_id and retrieve downloaded file paths.
- Send whitelisted image, video, txt/csv/pdf, or subtitle assets to a LAN bridge and let them become upload: parameters.
- When status is queued/running or doctor fails, notify the user and continue polling in a later independent call.
Best For
- Agent integration engineers calling multimodal generation workflows, who want schema-driven calls instead of hard-coded workflows.
- Engineers orchestrating local or LAN media generation, who need to manage bridge_url, asset uploads, and error feedback.
- Product engineers integrating whitelisted image, video, and subtitle assets into generation flows, with focus on upload limits and output paths.
- App maintainers debugging Comfypet run states, who need to distinguish queued, running, success, and warnings.
Related Skills
Excalidraw Wrap is an Excalidraw-focused wrapper, with tags for TypeScript, GitHub, and automation.
Calibrate vague brand inputs, expose contradictions, distill a brand core and positioning boundaries, then stress-test the result into an executable brand skeleton.
Generate localized Chinese brand names, naming directions, slogans, and risk checklists with reusable templates and trademark search reminders.
A beginner-friendly photo analysis tool that infers shooting parameters from visual features and suggests post-processing, optimization, and learning keywords.