The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to guide users through executing standard development commands such as uvicorn for local hosting and the Railway CLI for cloud deployment. These are appropriate for the skill's stated purpose of environment setup.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read state information from a local file named .wca-estado.json to determine the current progress of the onboarding.
Ingestion points: .wca-estado.json (local file system).
Boundary markers: The skill uses a specific schema logic (checking the fase field) to interpret the file content.
Capability inventory: The instructions involve guiding the user through file system reads, environment configuration, and deployment commands.
Sanitization: The skill relies on structured JSON fields (fase) rather than raw text interpolation for control flow, which limits the risk of injection from this file.

start