The Agent Skills Directory

[SAFE]: The skill operates as a structural auditing tool for manuscripts, performing local data validation and analysis without external dependencies or risky network operations.- [COMMAND_EXECUTION]: The skill uses the Bash tool to execute a bundled Python script, scripts/check_argument_governance.py. Analysis of this script confirms it is safe: it only imports standard libraries, restricts file access to the user-provided project root, and contains no logic for remote code execution or system manipulation.- [CREDENTIALS_UNSAFE]: The skill explicitly instructs the user to store API keys in environment variables rather than hardcoding them in files or logs, adhering to industry-standard security practices for secret management.- [PROMPT_INJECTION]: The skill manages risk from indirect prompt injection by enforcing a strict schema for its CSV data files and utilizing a deterministic checker script. It also includes instructions that prioritize local evidence over external model suggestions, reducing the impact of potentially malicious inputs in the evidence files.

argument-governance