themindshift-theme-approval-loop

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

• Third-party content exposure detected (high risk: 0.75). The skill’s “Required Reads” and “Loop Steps” explicitly require reading the active issue folder and “current external sources,” which at runtime can include outsider-authored free text (e.g., issue/discussion content or externally fetched signals) that the agent would ingest into the LLM context for theme candidate generation.