The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to check for and install the agent-spec CLI using cargo install agent-spec if the command is not found in the environment.
[COMMAND_EXECUTION]: The workflow involves executing agent-spec parse and agent-spec lint shell commands to validate the quality and syntax of generated .spec files.
[EXTERNAL_DOWNLOADS]: The agent is directed to fetch and install the agent-spec package from the public Rust crate registry (crates.io). This is identified as a vendor-provided resource for this skill.
[PROMPT_INJECTION]: The skill processes external task descriptions and BDD scenarios provided by users, creating an indirect prompt injection surface where malicious instructions could be embedded in data.
Ingestion points: Processes user-provided task requirements and reads local reference patterns from ./references/patterns.md.
Boundary markers: Utilizes structured Markdown headers (e.g., ## Intent, ## Constraints, ## Completion Criteria) to separate different types of data within the specification files.
Capability inventory: Possesses the ability to execute shell commands (via cargo and the agent-spec CLI) and read local files.
Sanitization: No specific input validation or sanitization of the user-provided data is mentioned before it is incorporated into the spec files and processed by the CLI tools.

agent-spec-authoring