mb-map-codebase
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from an existing repository to generate documentation. This creates a surface where malicious instructions embedded in the source code could influence the agent's behavior.
- Ingestion points: Repository files are scanned and parsed by the agent and its parallel subagents to generate final reports (SKILL.md).
- Boundary markers: The instructions lack explicit delimiters or "ignore" instructions for content found within the scanned repository files, although they mandate separating evidence from inferences (references/synthesis-checklist.md).
- Capability inventory: The skill possesses the capability to write multiple files to the local filesystem (in .memory-bank, .tasks, and .protocols directories) and spawn additional agent workers (SKILL.md).
- Sanitization: There is no mention of sanitizing, escaping, or validating repository content before it is processed or written into documentation.
Audit Metadata