The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform repository bootstrapping, file system checks, and validation tasks. It also reads directory listings from the user's home directory (~/.claude/skills/ and ~/.agents/skills/) to identify existing skills that might be relevant to the procedure being built.
[PROMPT_INJECTION]: The tool consumes a user-provided specification file and interpolates its content directly into generated CONTEXT.md and SKILL.md files. This creates a surface for indirect prompt injection where a malicious specification could inject harmful instructions into the resulting procedure artifacts. Ingestion points: The specification file path provided as the first positional argument ($1). Boundary markers: No specific delimiters or 'ignore embedded instructions' warnings are used when interpolating content into the procedure templates. Capability inventory: The generated procedure inherits the agent's capabilities to perform file system writes and execute shell commands. Sanitization: No input validation, escaping, or sanitization is performed on the text extracted from the specification file before it is written to the output files.

procedure-builder