agentleFS
Sign inStart free
Galbaz1Galbaz1/video-research-mcp/skills/qwen-image-integration/SKILL.md

qwen-image-integration

Skill23 starsChanged 40 days agoMITView on GitHub

Select the optional pinned external Qwen video-edit image tools for text generation, editing or image translation; check unsupported alpha intent and preserve source and result evidence before any authorized provider call.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the disabled optional integration documented in Qwen video-edit and its exact schema descriptor.

    line 14
  2. Very specific:

    Verify the pinned external source, selected private compatible runtime and grants.

    line 22
  3. Very specific:

    Read the actual qwen_image schema.

    line 28
  4. Distinctive:

    Keep paid calls disabled until durable intent/recovery guards are qualified.

    line 41
  5. Very specific:

    Keep task/model/request/reference/script/scene identity, synthetic label, anchors and continuation settings.

    line 42

≈ “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

Installed resources

For the Claude installer layout, read ../video-research-resources/docs/integrations/qwen-video-edit.md and ../video-research-resources/integrations/qwen/video-edit.json. The repository-relative links below apply in a source checkout or unpacked npm package. Installing these resources does not activate the external integration.

Use the disabled optional integration documented in Qwen video-edit and its exact schema descriptor. Source: QwenLM/Qwen-MM-Plugins at 07736672525443c7f8a3f6405eed37d2236f023f, package video-edit. This skill is independently authored against Apache-2.0 original wrapper/skill contracts; it copies no upstream code, assets or weights.

  1. Verify the pinned external source, selected private compatible runtime and grants. The entry is qwen-mm-plugins-video-edit or python -m qwen_mm_plugins_video_edit over stdio. Isolate QWEN_MM_CONFIG; keep QWEN_MM_NATIVE_MODE=true. Discovery needs no key and must not generate. Observed discovery is six tools, server version 1.1.0; native provider samples remain UNRUN because the tested runtimes lack DashScope/requests.
  2. Read the actual qwen_image schema. text_to_image and image_edit use qwen-image-2.0-pro; image_translate uses qwen-mt-image. Require a prompt for generation/editing, 1–3 images for editing and exactly one for translation. Preserve reference ordering and source hashes; reject overflow before upstream silently drops images. Translation uses language fields and ignores generation size/seed/prompt parameters.
  3. Reject alpha-required intent before submit: the schema has no transparent background field. Preserve the unsupported request in the durable receipt. PNG filenames and prompt words supply no alpha proof. Validate real provider size/count limits and reference availability before any spending.
  4. Obtain exact operation/source-submission/account/spend authority and verify selected origin and DASHSCOPE_API_KEY environment delivery. No install, upload or generation follows from readiness. The selected durable video route does not support these image modes. Keep paid calls disabled until durable intent/recovery guards are qualified.
  5. Keep task/model/request/reference/script/scene identity, synthetic label, anchors and continuation settings. Shared submit/SDK retries lack guaranteed idempotency; a timeout or unknown effect requires reconciliation, never a new generation call. Translation polls at 3s/120s and prints its task ID only on timeout, so retain provider identity outside the text formatter.
  6. Require bounded download, actual content hash and full decoded image metadata before accepting a final asset. URL-only output, upstream saved-path strings and request/URL-derived filenames are insufficient. Retain failed/expired URL evidence and unknown metadata. Durable receipts and final artifact acceptance remain separately unqualified.

Source mocks demonstrate contract behavior, not installed SDK availability, provider quality, durable recovery, alpha support or whole-feature acceptance.

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.

Check your setup, free