and-sweep

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

• Third-party content exposure detected (high risk: 0.70). The required workflow involves collecting workflow state and reading “comments or receipts” and other backend text at runtime (SKILL.md:105-110), which can include outsider-authored free text (e.g., issue/PR comments), so that content may be ingested into the agent’s LLM context for analysis/reporting.