mb-red-verify
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to analyze task implementation details, including code diffs, task plans, and evidence artifacts, which serves as an ingestion point for untrusted data and a potential surface for indirect prompt injection.\n- Ingestion points: The skill reads from files in
.protocols/<TASK_ID>/,.memory-bank/tasks/, and implementation diffs.\n- Boundary markers: There are no explicit instructions or delimiters defined to separate the untrusted input data from the agent's core instructions, nor are there commands to ignore embedded instructions within the task data.\n- Capability inventory: The skill is authorized to perform filesystem read and write operations to generate reports, update task indices, and file bug reports.\n- Sanitization: No specific sanitization, filtering, or validation logic is prescribed for the data ingested during the verification process.
Audit Metadata