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 as rg (ripgrep) and grep to 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 the docs/ 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
Risk Level
MEDIUM
Analyzed
Sep 7, 2026, 02:40 PM
Security Audit — agent-trust-hub — mathodology-project-orientation