mathodology-project-orientation
Warn
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a Python validation script (
.claude/skills/mathodology-dev-test-release/scripts/validate_repo.py) and shell utilities such asrg(ripgrep) andgrepto perform repository maintenance and boundary checks. - [REMOTE_CODE_EXECUTION]: The skill configuration mentions the use of
uvx free-search-mcp, which triggers the download and execution of an external Python package from a public registry at runtime. - [INDIRECT_PROMPT_INJECTION]: The skill processes repository metadata and file content, creating an attack surface where untrusted data could influence agent behavior.
- Ingestion points: Repository files including
README.md,AGENTS.md, and documents within thedocs/directory. - Boundary markers: The instructions do not define specific delimiters or instructions to the agent to ignore potential commands embedded in the files it reads.
- Capability inventory: The skill possesses the ability to execute shell commands and local Python scripts.
- Sanitization: The skill does not implement sanitization or validation of the contents read from the files before they are processed by the agent.
Audit Metadata