Przejdź do treści
Biblioteka skilli
Projektowanie i UX · Integracja z narzędziem

comfyui

/comfyui

Generuje obrazy i wideo AI przez workflow ComfyUI (FLUX, Stable Diffusion, modele wideo) na lokalnym GPU lub rozproszonej mocy Salad Cloud.

Repozytorium
agentbox ↗
Licencja
AGPL-3.0
Ostatnia zmiana repozytorium
27.10.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
28
Skrypty
5
Komentarze HTML
0
Adresy URL w SKILL.md
4
Przestarzałość

Brak sygnałów. Repozytorium zmienione 27.10.2026, bez odwołań do wycofanych modeli i API.

Licencja

AGPL-3.0, copyleft: przeróbki udostępniane innym muszą mieć tę samą licencję.

Co sprawdzić przed instalacją
  • 5 skryptów w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.

Skrypty: mcp-server/server.js, mcp-server/server.test.js, scripts/comfy_client.py, scripts/download_video_models.py, scripts/tests/test_comfy_client.py

Co robi i kiedy

Generuje obrazy i wideo AI przez workflow ComfyUI (FLUX, Stable Diffusion, modele wideo) na lokalnym GPU lub rozproszonej mocy Salad Cloud.

Oryginalny opis (EN)

Generate AI images and video with ComfyUI's node-based workflows (FLUX, Stable Diffusion, video models) on a local GPU or distributed Salad Cloud compute. Use when a task needs text-to-image, image-to-image, or text/image-to-video generation, ComfyUI workflow authoring, or scaling generation across GPUs.

Jak działa

  1. Container architecture (read first)
  2. Quick path
  3. Capabilities
  4. When to use this skill
  5. When not to use
  6. References
  7. Integration with other skills
  8. Related skills
Sekcje z nagłówków SKILL.md, w oryginale · 139 linii, 853 słów
Podgląd SKILL.md
---
name: comfyui
description: "Generate AI images and video with ComfyUI's node-based workflows (FLUX, Stable Diffusion, video models) on a local GPU or distributed Salad Cloud compute. Use when a task needs text-to-image, image-to-image, or text/image-to-video generation, ComfyUI workflow authoring, or scaling generation across GPUs."
depends_on_mcps:
  - comfyui
---

# ComfyUI Skill

Drive ComfyUI for AI image/video generation, node-based workflow authoring, and
distributed GPU compute via the Salad Cloud API. ComfyUI runs in an **external
Docker container** reached over the Docker network — you talk to it through its
HTTP API, never the local filesystem.

`agentbox.toml`'s `skills.media.comfyui_builtin = false` and this skill's
`mcp.json` entry `"x-agentbox-gate": "env:ENABLE_COMFYUI_EXTERNAL"` name the
same fact: ComfyUI is not baked into the image, it runs externally.

## Container architecture (read first)

ComfyUI runs in an EXTERNAL Docker container, not locally. When calling from Claude
Code (its own container), pick the endpoint that matches where you are:

| Scenario | Endpoint | Notes |
|----------|----------|-------|
| From Claude Code container | `http://comfyui:8188` | Docker network hostname |
| From host machine | `http://localhost:8188` | Port exposed to host |
| Container IP (fallback) | `http://172.18.0.X:8188` | Check with `ping comfyui` |

Networking rules that actually bite:
1. From inside the Claude Code container, `localhost:8188` will not reach ComfyUI —
   use the `comfyui` Docker hostname.
2. You cannot read the container filesystem; retrieve outputs via
   `/view?filename=...&type=output`, not filesystem paths.
3. The comfyui container's volume mounts (output/input/models under
   `/mnt/mldata/.../comfyui/...`) are host-accessible only, not from here.

### Check container status

From the Claude Code container, use the HTTP health check — this is the canonical
liveness probe and needs no Docker access:
```bash
ping -c1 comfyui  # Should show IP like 172.18.0.X
curl -s http://comfyui:8188/system_stats | jq '.devices[0].name'
```

> **Docker-daemon diagnostics (no `sudo`).** agentbox mounts the host Docker socket,
> so plain `docker` talks to the host daemon from inside the container. Do **not**
> prefix with `sudo` — it is blocked by `no-new-privileges` and will fail:
> ```bash
> docker ps --filter "name=comfyui"
> docker logs comfyui --tail 20
> ```
> Prefer the `comfyui:8188` HTTP check above; only fall back to these when the HTTP
> probe is inconclusive.

## Quick path

To build or repair the local GPU service, follow
[`references/sidecar-operations.md`](references/sidecar-operations.md).

For resumable video jobs, use `scripts/comfy_client.py`; read
[`references/stopmotion-proof.md`](references/stopmotion-proof.md) for the Blender
endpoint and clay-LoRA comparison, and
[`references/agent-kit.md`](references/agent-kit.md) for graph validation and durable
submission receipts. Local video model setup and workflows are documented in
[`references/video-models.md`](references/video-models.md). For a complete explainer
of the current repository, start with `codebase-video`.

For a new task, inspect live `/system_stats`, `/queue` and the relevant
`/object_info/<node>` schemas. Installed weights are discovered from loader
choices, never from a static table in this skill. The bundled FLUX example
requires separate provisioning and is not the current installed default.

Use native HTTP for uploads, durable submission/history and media retrieval;
WebSocket events provide responsive progress but do not establish completion.
The MCP adapter exposes schema discovery, queue/status, model listing and safe
idle-model unloading. It accepts API graphs rather than guessing a model from
an old convenience recipe. See [control planes](references/control-planes.md)
for the official Comfy MCP alternative and its remote-filesystem limitations.

For model preparation, `scripts/download_video_models.py --campaign core --plan`
reports shared components and storage without inference. The pinned manifest
includes FIBO2, Klein, LTX2.5 layout/style and H3 control, with larger comparisons
separate. Follow the [preparation guide](../../docs/developer/media-model-preparation.md)
for rebuilds and independent CPU-only provisioning.

## Capabilities

- Text-to-image (text2img) and image-to-image (img2img) generation
- Author and execute node-based workflows programmatically
- Video generation (AnimateDiff, CogVideoX, HunyuanVideo, LTX, Mochi, WAN, Cosmos)
- Manage models, LoRAs, checkpoints, VAEs
- Upscaling and post-processing
- Deploy to Salad Cloud for distributed GPU compute at scale

## When to use this skill

- Generate AI images from text descriptions or transform image-to-image
- Design ComfyUI workflows programmatically
- Batch-process generation tasks
- Use or fine-tune LoRA models with FLUX/SD
- Generate AI video from text or images
- Deploy image generation at scale across distributed GPUs

## When not to use

- 2D image manipulation (resize, crop, format convert) — use the **imagemagick** skill
- 3D modelling and scene creation — use the **blender** skill
- Video transcoding, editing, or audio extraction — use the **ffmpeg-processing** skill
- Diagrams, flowcharts, architecture visuals — use the **mermaid-diagrams** skill
- ML model training (classification, NLP, time series) — use **pytorch-ml** or **flow-nexus-neural**

## References

Detail lives in `references/` — pull the file that matches the task:

- [`references/flux2-quickstart.md`](references/flux2-quickstart.md) — full FLUX 2
  generate/monitor/download flow, one-liner, VRAM management, current model + node tables.
- [`references/api-and-parameters.md`](references/api-and-parameters.md) — local server
  operations, API endpoints, text2img parameter table, sampler/scheduler lists, output
  conversion, webhooks, error handling, directory layout.
- [`references/workflows.md`](references/workflows.md) — FLUX 2 and legacy FLUX 1 workflow
  JSON structures, supported image/video models, batch-generation Python.
- [`references/salad-cloud.md`](references/salad-cloud.md) — Salad SDK deployment, the full
  recipe catalog (image/video/LLM), hardware recommendations, benchmarks, performance notes.

Bundled runnable assets in this directory:
- `comfyui-generate` — FLUX 2 one-shot generator (quick path above), on PATH
  from the `agentbox-ops` crate.
- `anima_workflow.json` / `anima_import_workflow.json` — working sample animation
  workflow graphs, kept as reference structures for `generate.py`.
- `mcp-server/` — ComfyUI MCP server surface.

## Integration with other skills

- **imagemagick** — post-process generated images
- **ffmpeg-processing** — assemble/transcode generated video
- **blender** — 3D-to-2D workflows
- **pytorch-ml** — custom model training

## Related skills

- `art` for editorial illustration, infographics and comics via the Nano Banana image models; ComfyUI is the local-GPU FLUX/Stable Diffusion/video path.
Pliki w paczce (28)
  • SKILL.md 7 KB
  • anima_import_workflow.json 4 KB
  • anima_workflow.json 2 KB
  • mcp-server/package-lock.json 79 KB
  • mcp-server/package.json 1 KB
  • mcp-server/server.js 15 KB
  • mcp-server/server.test.js 1 KB
  • models/stopmotion-candidates.json 26 KB
  • references/agent-kit.md 4 KB
  • references/api-and-parameters.md 4 KB
  • references/control-planes.md 2 KB
  • references/flux2-quickstart.md 7 KB
  • references/h3-compatibility.md 4 KB
  • references/ltx23-models.md 5 KB
  • references/salad-cloud.md 6 KB
  • references/sidecar-operations.md 4 KB
  • references/stopmotion-proof.md 2 KB
  • references/video-models.md 6 KB
  • references/workflows.md 4 KB
  • scripts/comfy_client.py 17 KB
  • scripts/download_video_models.py 14 KB
  • scripts/tests/test_comfy_client.py 11 KB
  • workflows/ltx-2.3-t2v-smoke.api.json 4 KB
  • workflows/ltx-2.3-t2v.api.json 4 KB
  • workflows/minimax-h3-i2v.api.json 3 KB
  • workflows/minimax-h3-r2v.api.json 3 KB
  • workflows/minimax-h3-t2v-smoke.api.json 3 KB
  • workflows/minimax-h3-t2v.api.json 3 KB

Wersje

RepozytoriumWynikLicencja
Ta strona · dreamlab-ai/agentbox100AGPL-3.0
mrsandipmandal/CineForge100Apache-2.0
TerminalSkills/skills100Apache-2.0
calesthio/OpenMontage100AGPL-3.0
Zulgo/video-generation-automation100AGPL-3.0
G1Joshi/Agent-Skills100MIT
SlavaSexton/ComfyUI-Agent-Kit95Apache-2.0

Migawka bazy z 13.11.2026 · wersja paczki 8c1309dc52b7 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100comfy
    Generuje obrazy, wideo, audio i modele 3D w Comfy Cloud: wyszukuje modele i szablony workflow, uruchamia własne workflow ComfyUI i zarządza zadaniami przez hostowany serwer MCP.
  • 90stable-diffusion-image-generation
    Generuje obrazy z opisów tekstowych modelami Stable Diffusion przez Diffusers: przekształcenia obraz-obraz, inpainting i własne potoki.
  • 100cloudgpu-comfyui-images
    Wynajmuje GPU do generowania obrazów w ComfyUI modelem FLUX, gdy lokalna karta nie wystarcza.
  • 100diffusion-models-stable-diffusion-flux
    Generuje i edytuje obrazy modelami dyfuzyjnymi, np. Stable Diffusion czy Flux, z negatywnymi promptami i sterowaniem ControlNet.
  • 90stable-diffusion-image-generation
    Generuje obrazy z tekstu, przekształca obraz w obraz, maluje ubytki i buduje własne potoki diffusion w Stable Diffusion przez HuggingFace Diffusers.
  • 100flux-image
    Generuje i edytuje obrazy Flux oraz wideo FLUX 3 przez API AceDataCloud: obrazy z promptów tekstowych, edycja istniejących obrazów instrukcjami i tymczasowe szkice wideo.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.