The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documentation describes a workflow involving the ingestion of external data via scenario files and agent source code, which serves as an input surface for the simulation's LLM judge.
Ingestion points: Agent source code and scenarios.yaml (referenced in SKILL.md).
Boundary markers: None mentioned.
Capability inventory: Shell command execution via lk agent simulate (referenced in SKILL.md).
Sanitization: None mentioned.
[SAFE]: The skill provides legitimate documentation for the lk (LiveKit) CLI tool. It describes standard development workflows for testing real-time voice and text agents, including iterative testing and regression testing in CI/CD environments.
[SAFE]: The documentation includes a safety disclaimer noting that the CLI tool asks for user confirmation before performing actions like code uploads for scenario generation.
[SAFE]: The use of export to JSON and list subcommands for triage follows best practices for managing build artifacts and diagnostic logs in a development pipeline.

running-livekit-simulations