The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes input from source code diffs and logs to build its test plan. 1. Ingestion points: Source code changes from git diff and git log, plus project metadata from files like package.json. 2. Boundary markers: The skill relies on standard Markdown structure without specialized delimiters to segregate analyzed content from instructions. 3. Capability inventory: The agent uses Bash, Read, and Write tools to interact with the repository and execute local testing tools. 4. Sanitization: No explicit sanitization of code-derived text is performed before it enters the agent's context.
[DYNAMIC_EXECUTION]: The skill constructs and executes database queries and shell commands dynamically based on the project environment. 1. Patterns: It builds a DB_CMD string from local environment variables and project-specific CLI tools to perform read-only introspection. 2. Context: These actions are intrinsic to the tool's primary purpose of verifying code changes and are limited to local or verified containerized services.

qakit