The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses dynamic context injection (the !command syntax) within its script templates (e.g., scripts/create-msw-handler.md, scripts/create-test-case.md) to execute shell commands such as grep, find, and date at runtime. These are used to automatically detect project frameworks and configuration details.
[DATA_EXFILTRATION]: A command in scripts/create-msw-handler.md explicitly searches through sensitive environment files (grep -r "API_URL\|VITE_API\|NEXT_PUBLIC_API" .env*) to extract API base URLs. Accessing .env files is a security concern as they frequently contain secrets, keys, and credentials that should not be exposed to the agent or its context.
[PROMPT_INJECTION]: Indirect prompt injection vulnerability surface detected.
Ingestion points: User-supplied $ARGUMENTS in script files (scripts/create-msw-handler.md, scripts/create-test-case.md, scripts/create-test-fixture.md).
Boundary markers: Absent; user input is directly interpolated into markdown and code templates.
Capability inventory: The agent has access to Read, Glob, Grep, WebFetch, and WebSearch tools, and executes shell commands via dynamic context.
Sanitization: No sanitization, escaping, or validation is performed on the user-provided arguments before interpolation.

testing-unit