plan-orchestrate

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

• Third-party content exposure detected (high risk: 0.75). The required workflow reads the user-supplied <plan-doc-path> and then emits step task descriptions that inherit plan text verbatim (including any outsider-authored content inside that document), which becomes readable free text in the LLM context via the skill’s own decomposition/compression pipeline.