code-to-diagram
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script scripts/analyze_codebase.py executes the external command mmdc (Mermaid CLI) using subprocess.run to render diagrams. While the command is constructed using a list of arguments minimizing shell injection risks, it relies on an external binary being present on the system path.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code files to extract dependency information and generate diagrams. 1. Ingestion points: The script reads file contents in scripts/analyze_codebase.py from any directory provided by the user. 2. Boundary markers: The generated Mermaid output and JSON results do not include explicit security boundary markers or instructions for the agent to ignore embedded content. 3. Capability inventory: The script has file read access to the target codebase and file write access for outputting diagrams, along with the ability to execute the mmdc command. 4. Sanitization: While Mermaid identifiers are sanitized using alphanumeric filtering, file names and directory structures are included as labels in the output without specific escaping for potential prompt injection patterns.
Audit Metadata