The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process code review feedback from external sources, which represents an ingestion surface for potentially malicious or misleading instructions.
Ingestion points: External code review comments received via chat or GitHub pull request threads.
Boundary markers: No explicit data delimiters (like triple backticks or XML tags) are specified for the input, but the skill implements a logical boundary by requiring the agent to "verify before implementing" and "check against codebase reality."
Capability inventory: The skill utilizes the grep command for searching the codebase and the gh (GitHub) CLI for interacting with the GitHub API to post replies.
Sanitization: There is no explicit instruction to sanitize Pull Request identifiers or repository paths used in the gh api call templates, relying on the environment to provide valid context.
[COMMAND_EXECUTION]: The skill instructs the agent to use standard development tools to perform its tasks.
Evidence: The use of grep is prescribed for YAGNI (You Ain't Gonna Need It) checks to verify if suggested features are actually used in the codebase.
Evidence: The skill provides a specific command template gh api repos/{owner}/{repo}/pulls/{pr}/comments/{id}/replies for responding to GitHub reviews. These tools are used for their intended purpose in a software development workflow.

receiving-code-review