odata-vocabularies-sync

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to take the provided vocabulary URL and insert it verbatim into source files (e.g., as the uri value), which would force the LLM to handle and output any secret contained in that URL (such as tokens in query strings) directly.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow can fetch and ingest outsider-provided JSON from a user-supplied URL argument (must end in .json) to extract namespace and $Alias, then later the update tool fetches all registered vocabularies including that newly added one.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs fetching and ingesting external vocabulary JSON at runtime (e.g. https://sap.github.io/odata-vocabularies/vocabularies/SomeNew.json) and the update tool also fetches registered vocabulary JSONs to generate/update resource TypeScript files, so remote content directly controls generated code/resources.