github

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

• Third-party content exposure detected (high risk: 0.75). The required workflow uses the gh CLI to fetch PR/issue/run details (e.g., gh pr view ... --json ... --jq ... and gh run view ...), which at runtime can ingest outsider-authored free text like PR/issue bodies and CI log text into the agent’s LLM context via the command output.