analyze-memory-snapshot

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

• Third-party content exposure detected (medium risk: 0.30). The workflow reads a user-provided snapshot file (JSON or pickle) from an arbitrary path at runtime (e.g., load_snapshot() → json.loads(path.read_text()) / pickle.load(handle), and the snapshot’s string fields (e.g., frames[].filename/name and device_traces[].action) are then formatted into the LLM-visible output.