pseudocode
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for the agent to ingest and analyze external source code files to map architecture and call flows. This introduces a vulnerability surface where malicious instructions embedded in the code being analyzed (e.g., in comments or string literals) could influence the agent's behavior.
- Ingestion points: The skill instructions in Section 2 ('Read the parts') direct the agent to 'opening each file on the path' and 'For every function reached, record its declaration... its signature... the effects it causes'.
- Boundary markers: The instructions lack explicit delimiters or 'ignore embedded instructions' directives to prevent the agent from following commands found within the data files it is processing.
- Capability inventory: The skill allows for recursive file system reads across a codebase and includes a 'save' capability where the agent can write the final render to a new file at the user's request.
- Sanitization: There are no requirements for sanitizing, escaping, or validating the content read from the source files before it is processed by the LLM or rendered into the final output.
Audit Metadata