The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing the agent-spec CLI tool via cargo install. This is a functional dependency required for the tool's core operations of parsing and linting specifications.- [COMMAND_EXECUTION]: The skill instructs the agent to execute CLI commands such as agent-spec parse, agent-spec lint, and agent-spec init. These are legitimate operations used for scaffolding, validating, and ensuring the quality of the task specifications.- [PROMPT_INJECTION]: The skill manages a surface for indirect prompt injection through its authoring workflow. (1) Ingestion points: User-provided task descriptions and intent. (2) Boundary markers: The skill enforces strict section headers (e.g., ## Intent, ## Boundaries) and path globs for Allowed Changes. (3) Capability inventory: Subprocess calls to the agent-spec CLI for parsing and validation. (4) Sanitization: The skill implements a mandatory quality gate using agent-spec lint with a minimum score of 0.7 to ensure that generated specs are deterministic and testable.

agent-spec-authoring