The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of Atmos CLI commands (e.g., atmos casts setup, atmos cast render) to perform its primary tasks, which grants the agent interaction capabilities with the underlying system.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from Atmos repositories and fixtures to generate documentation artifacts. This creates a surface where malicious repository content could potentially influence the agent's behavior.
Ingestion points: Atmos repository files, YAML configurations, and demo fixtures under demo/casts.
Boundary markers: The skill does not define specific delimiters for untrusted content, though it does provide manual review guidelines.
Capability inventory: The skill uses the Atmos CLI to run workflows and writes output to the website/static/casts directory.
Sanitization: Instructions are provided for the user/agent to review plain text recordings for secrets and unstable data before commitment.
[DYNAMIC_EXECUTION]: The skill uses custom YAML tags (!include) in its configuration examples to dynamically load content from external files like cast-defaults.yaml.

atmos-asciicast