The Agent Skills Directory

Command Execution: The skill instructs the agent to use standard development tools such as git and the GitHub CLI (gh). These tools are used for repository maintenance tasks like auditing staged changes and managing file states.
Instructional Directives: The skill contains specific instructions to "override" default rules regarding the staging of files. In the context of this agent's task, this is a functional directive to ensure that technical fixes identified during the critique phase are correctly applied to the repository using git add.
Security Auditing Instructions: The skill includes a 'Security & Payload Awareness' section that explicitly directs the agent to scan for and reject malicious patterns in the code it reviews, such as prompt injection, data exfiltration, and unauthorized command execution. This acts as a defensive layer for the repository.
Indirect Prompt Injection Surface: The skill processes untrusted data in the form of repository code and scripts. To mitigate risks, it provides a technical checklist and sanitization instructions (e.g., scanning for 'ignore all rules' strings) to prevent malicious code from being approved or executed during the auditing process.
Vendor Integration: The skill references gemini-cli and internal paths like tools/gemini-cli-bot/, which are consistent with the vendor's own infrastructure and tools for automated repository management.

critique