The Agent Skills Directory

[SAFE]: The skill provides structured guidance for code review output routing and formatting. It implements logical branching based on the execution context to ensure findings are placed in the correct destination (local file or GitHub comment) without introducing security vulnerabilities.\n- [COMMAND_EXECUTION]: References to tools like gh pr comment and Model Context Protocol (MCP) comment update tools are appropriate for the stated purpose of posting pull request summaries. The skill follows best practices by restricting specific tools to ensure workflow consistency.\n- [DATA_EXPOSURE]: File operations are restricted to writing summaries to predefined local or temporary files such as review-summary.md or /tmp/review-summary.md. No patterns of sensitive data harvesting or unauthorized network transmission were identified.

posting-review-summary