The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses git diff to identify modified files in the repository and grep to detect specific keywords related to databases and external services (e.g., Prisma, transaction, fetch, axios). This information is used to route files to the appropriate review agents.
[DYNAMIC_EXECUTION]: The orchestrator dynamically invokes sub-agents (e.g., review-comments, review-acid) using the platform's parallel execution capability. It constructs prompts for these sub-agents to process specific files and return structured results based on a predefined schema.
[INDIRECT_PROMPT_INJECTION]: As a code review tool, the skill ingests untrusted source code provided by the user. This creates a surface for indirect prompt injection if the source code contains malicious instructions intended for the review agents. The risk is mitigated by the use of structured output schemas and instructions that guide the sub-agents to use file-reading tools rather than direct string interpolation.

review-all