The Agent Skills Directory

[SAFE]: The skill functions as a procedural manual for code quality assurance. It includes specific checks for security best practices, such as preventing the inclusion of hardcoded secrets and ensuring the use of parameterized SQL queries.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted source code from pull requests and files (Ingestion points). The agent possesses capabilities including Read, Grep, Glob, and Bash (Capability inventory). While the skill lacks explicit boundary markers or sanitization for this data, it includes a safety constraint directing the agent to use pre-existing test artifacts rather than executing the code directly.

code-review