The Agent Skills Directory

[SAFE]: The skill operates as a compliance reviewer and repository auditor. All automated checks and runtime logic are implemented via local scripts (run_checks.sh, repo_review_suite.mjs, agent_runner.mjs) included in the skill's distribution. No external network requests, obfuscation, or unauthorized data access patterns were identified during the analysis.
[COMMAND_EXECUTION]: The skill executes local scripts to perform file analysis and process management. For example, run_checks.sh utilizes standard utilities (grep, sed, wc) to verify skill frontmatter and structure, while Node.js scripts manage state transitions and test execution. These actions are aligned with the skill's primary purpose as a development tool.
[SAFE]: External agent interaction is handled via the agent_runner.mjs script, which spawns local CLI processes. The configuration for these tools (found in agent_registry.json) uses standard non-interactive flags. The workflow includes a mandatory 'Critical Verification' step where the host agent must independently verify all suggestions, providing a layer of human-in-the-loop oversight.

ln-162-skill-reviewer