github-goal-issue-triage

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

• Third-party content exposure detected (high risk: 0.95). The required workflow fetches outsider-authored repository content at runtime—specifically GOAL.md from each repo (and then open issue titles/bodies/comments) via the GitHub API/connector—which becomes readable text in the LLM context for summarization/triage.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches each repository's GOAL.md at runtime and uses its contents to drive triage decisions (e.g., repositories provided as "https://github.com/owner/name" or "git@github.com:owner/name.git" are fetched and GOAL.md is treated as the required "north star"), so remote repository content controls the agent's behavior.