The Agent Skills Directory

Command Execution: The skill utilizes local Python scripts that invoke system commands for repository management and configuration verification. Specifically, model_pr.py uses the subprocess module to coordinate git and GitHub CLI operations. These commands are executed using argument lists, which is a standard practice to prevent shell injection, and the skill is designed to require user oversight for all external actions.\n- Indirect Prompt Injection Surface: The skill is designed to process external repository documentation, including AGENTS.md and SECURITY.md files. There is a potential security consideration that these documents could contain instructions intended to influence the agent's behavior. The skill addresses this by including explicit instructions to treat all external content strictly as data and to prioritize human-led verification of all proposed changes.

model-verify