The Agent Skills Directory

[SAFE]: Analysis of the skill files revealed no malicious code, obfuscation, or exfiltration vectors. The content is focused on legitimate development and testing practices.
[NO_CODE]: The skill is a collection of Markdown documentation and code examples; it does not contain any executable scripts or binary files.
[PROMPT_INJECTION]: An indirect prompt injection surface is present as the agent is instructed to process external requirements (PRDs) and application code to generate and run tests. Ingestion points: references/planner-agent.md and references/generator-agent.md. Boundary markers: Absent. Capability inventory: Execution of local test suites (Playwright/pytest). Sanitization: The skill recommends mocking and validation patterns but does not specify input sanitization for PRDs.
[COMMAND_EXECUTION]: The skill provides instructions for agents to execute commands for running tests and starting local application instances.
[EXTERNAL_DOWNLOADS]: References installation of official evaluation packages (deepeval, ragas) from standard registries.

testing-llm