The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of the ralph CLI tool to perform its primary functions, including ralph hats validate, ralph hats graph, ralph hats show, and ralph run as described in references/commands.md and SKILL.md.
[PROMPT_INJECTION]: Identifies an indirect prompt injection surface where the skill processes user-authored YAML configuration files. 1. Ingestion points: The skill reads existing configuration files from .ralph/hats/*.yml as part of the inspection and improvement workflows. 2. Boundary markers: There are no explicit boundary markers or instructions to the LLM to ignore embedded commands within the ingested YAML data. 3. Capability inventory: The skill possesses the capability to write to the local file system and execute subprocesses via the ralph CLI. 4. Sanitization: No evidence of sanitization or filtering of the natural language fields within the YAML files is present.

ralph-hats