The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a variety of standard development tools including npm, pnpm, uv, pytest, ruff, and pyright. These are used to verify the build, types, and quality of the local codebase. These operations are core to the skill's intended purpose as a verification loop.
[CREDENTIALS_UNSAFE]: The skill implements a security scan phase that uses grep to identify potential hardcoded secrets such as sk- or api_key within the project files. This is a defensive security feature designed to help users identify and remove sensitive credentials before committing code.
[PROMPT_INJECTION]: The skill processes and displays output from local tools and file content which represents a surface for indirect prompt injection.
Ingestion points: Tool outputs from build logs, test results, linting reports, and git diffs are displayed to the agent.
Boundary markers: The skill does not define specific boundary markers or instructions for the agent to ignore embedded instructions within tool outputs.
Capability inventory: The skill has the capability to execute shell commands, read files via grep, and perform git operations.
Sanitization: No sanitization or filtering of tool output is performed before it is presented to the agent.

verification-loop