source-to-system

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

• Third-party content exposure detected (high risk: 0.75). This skill’s “Capture” phase explicitly instructs reading and preserving the root source and its dependencies (including external/untrusted material) into SCRATCH, which would then be available as readable text for the agent’s LLM context—i.e., outsider-authored free text from runtime-fetched or user-provided external sources.