RHDH plugin for fetching and extracting OCI skills images build via 'skillctl' CLI

[gabemontero]

Reference JIRA: https://redhat.atlassian.net/browse/RHIDP-16943

Description

build a typescript plugin that via code can fetch and process OCI skills images built by https://github.com/redhat-et/skillimageConnect your Github account

As a starting reference, see if the nodejs based image client at https://github.com/oras-project/rust-oci-client-nodeConnect your Github account can serve as either an imported library or a reference implementation that can be copy/pasted into the new RHDH plugin build

Acceptance Criteria

• running the plugin via the ‘yarn dev’ or ‘yarn start’ command in your IDE, create a plugin that takes image refs as configuration, and fetches the configured images

• confirm if the image meets the format specified by https://github.com/redhat-et/skillimage

  • one such image is quay.io/gabemontero/hello-world-skill:1.0.0-draft
  • golang code from skillimage that inspects images is at https://github.com/redhat-et/skillimage/blob/main/pkg/oci/inspect.go
  • typescript code that pulls images is shown at https://github.com/oras-project/rust-oci-client-node#usage

• extract the skillimage.yaml and SKILLS.md files onto local storage,

  • see https://github.com/redhat-developer/rhdh-plugins/blob/main/workspaces/ai-integrations/plugins/catalog-techdoc-url-reader-backend/src/plugin.ts#L40 and https://github.com/redhat-developer/rhdh-plugins/blob/main/workspaces/ai-integrations/plugins/catalog-techdoc-url-reader-backend/src/plugin.ts#L68-L99 for examples of how to store files locally form a RHDH plugin

[fullsend-ai-triage]

🤖 Finished Triage · ✅ Success · Started 7:16 PM UTC · Completed 7:23 PM UTC

Commit: 2c08c10 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $1.06

[fullsend-ai-triage]

Related work: Issues #4047–#4050 in workspace/boost cover OCI Skill Registry connectors (catalog entity providers using OCI Distribution Spec). They share the OCI skill image domain but target a different workspace, architecture (entity emission vs. file extraction), and JIRA epic. Worth coordinating to avoid duplicated OCI client logic — the shared boost-connector-utils package (#4039, now closed/merged) or a similar shared utility could be reused.

Proposed test case:

// 1. Image fetch – mock OCI client returns valid manifest + blobs
//    Assert: plugin downloads and reassembles layers correctly

// 2. Format validation – valid skillimage passes; image missing
//    skillimage.yaml or SKILLS.md layers fails with descriptive error

// 3. File extraction – assert skillimage.yaml and SKILLS.md are
//    written to temp dir following catalog-techdoc-url-reader-backend
//    pattern (fs.promises.mkdtemp + fs.promises.writeFile)

// 4. Config handling – plugin reads image refs from app-config.yaml;
//    missing or invalid config yields clear error, not a crash

---

Labels: New feature request for building an OCI skills image plugin; 'feature' and 'enhancement' labels match repo conventions for new functionality.

---

Next steps:

• /fs-code — agent creates a PR to implement this issue
• /fs-code <your instruction> — agent implements with your specific guidance

[gabemontero]

/fs-code

[fullsend-ai-coder]

🤖 Finished Code · ✅ Success · Started 7:26 PM UTC · Completed 7:46 PM UTC

Commit: 2c08c10 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Cost: $4.20