The Agent Skills Directory

Indirect Prompt Injection Surface: The skill is designed to have an agent read and analyze test files and production source code within a repository. This creates a surface where instructions embedded within those files (such as in comments or test strings) could potentially influence the agent's behavior during the audit process.
Ingestion Points: The agent reads various test files, production source code, and configuration files (e.g., AGENTS.md, vitest.config.ts, package.json).
Capabilities: The skill utilizes shell commands via pnpm and vitest to run tests and baseline the environment. It also facilitates the deletion of test files and production code deemed to be 'junk'.
Boundary Markers: There are no specific boundary markers or 'ignore' instructions defined for the data ingestion process.
Sanitization: The instructions do not specify sanitization or validation steps for the content of the files being audited.
Command Execution via Local Tooling: The skill involves executing shell commands such as pnpm, vitest, and git. These commands are used for baseline discovery, validation of changes, and landing PRs. While these are standard development practices, they represent a significant capability within the agent's environment.
External Reference to GitHub: The skill references an external resource on GitHub (github.com/openclaw/openclaw) as a source for its logic. This is a well-known service and the reference is documented as a source for the skill's adaptation.

test-audit