context-map
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill incorporates a bundled Python script,
scripts/build-context-map.py, which is intended to be executed by the agent to perform local codebase analysis. The script uses standard libraries to search for text patterns in files and identify dependencies, limiting its operations to local filesystem reads and reporting via standard output. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and summarizes content from arbitrary local project files during its mapping process.
- Ingestion points: Project files are read using
path.read_text()within thescripts/build-context-map.pyscript. - Boundary markers: There are no explicit boundary markers or instructions to ignore embedded content in the generated output.
- Capability inventory: The script is restricted to read-only access of local files; it lacks capabilities for network communication, file system modification, or secondary command execution.
- Sanitization: The script performs no sanitization of the content it reads, but the overall risk is low due to the restricted environment and the tool's specific purpose.
Audit Metadata