The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and implement feedback from external code reviews. Because this feedback is untrusted data, it serves as a potential surface for indirect prompt injection. The skill partially mitigates this risk by explicitly instructing the agent to verify all suggestions against the codebase and to push back on technically unsound requests.- [COMMAND_EXECUTION]: The instructions direct the agent to use the grep command to search the local codebase to verify the usage of specific endpoints or logic before following reviewer suggestions.- [NETWORK_OPERATION]: The skill uses the gh api tool to communicate with GitHub for replying to pull request comments. This is a standard operation for this use case and targets a well-known service, posing no significant risk.

receiving-code-review