The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run shell commands (e.g., rg) to inventory files and verify repository scripts. It incorporates safety constraints, explicitly directing the agent not to install dependencies, mutate external systems, or run destructive commands during the verification phase.\n- [SAFE]: The skill does not contain any hardcoded credentials, obfuscated code, or network operations to external domains. Its behavior is limited to local repository analysis and documentation generation.\n- [PROMPT_INJECTION]: While the skill involves reading untrusted repository data (like READMEs) which could contain indirect prompt injections, it mandates a verification step where the agent must reconcile claims against actual repository evidence (manifests, scripts, and CI workflows) to ensure accuracy and safety.

agent-instructions