diagnosing-bugs

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

• Third-party content exposure detected (medium risk: 0.30). Source: the skill’s required workflow in SKILL.md; runtime path: the agent builds and runs a “tight feedback loop” using user-provided commands/outputs and may ingest pasted error messages or HAR/log dumps during Phase 1/2, which are outsider-authored free text supplied via the user.