github-a11y-scanner

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs the agent to query, fetch and parse GitHub Issues and related data from public repositories (see "Parsing Scanner-Created Issues" and the search queries like "repo:{REPO} is:issue" and cross-repository discovery), which are untrusted, user-generated third‑party contents that the agent is expected to interpret and that directly influence severity, confidence, and follow-up actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill references and relies on the GitHub Action "github/accessibility-scanner" (https://github.com/github/accessibility-scanner or uses: github/accessibility-scanner@v2), which is fetched and executed at workflow runtime and therefore constitutes a remote code dependency run during execution.