The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute hoyeon-cli, a project-specific command-line interface, for initializing specifications, merging data via stdin, and validating project layers.
[COMMAND_EXECUTION]: It generates automated hooks in .claude/settings.json that execute formatters (e.g., Prettier, ESLint, Ruff) and linters upon file modification. It also includes a security-positive hook to block direct edits to .env files using grep and exit 1.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) as it processes untrusted user input to derive project architecture and requirements.
Ingestion points: User input is ingested via AskUserQuestion during the Architecture Interview (L2) in SKILL.md.
Boundary markers: The skill uses a structured spec.json format and enforces schema validation via hoyeon-cli spec validate at every layer transition.
Capability inventory: The skill has access to Bash (command execution), Write (file modification), and Task (subagent spawning) across its entire workflow.
Sanitization: Input is validated against a schema (--schema v1) and undergoes an automated 'L2 architecture review' by a general-purpose subagent to ensure coherence and quality before execution.

scaffold