cadence-adherence-telemetry

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

• Third-party content exposure detected (high risk: 0.75). High probability: the required workflow reads local context/ files (outsider-authored record bodies/issues/PR-like content could be present there), then render_cadence_report.py injects fields like f.human_readable/record details into the generated markdown that is persisted and surfaced in the agent’s LLM-visible text.