The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its core functionality of processing external design documents. 1. Ingestion points: The skill reads files from 'design/gdd/', 'docs/engine-reference/', and 'docs/architecture/'. 2. Boundary markers: The instructions do not employ delimiters or specific 'ignore' directives to isolate ingested content. 3. Capability inventory: The skill is permitted to use the Bash, Write, and Task tools. 4. Sanitization: No sanitization or structural validation is applied to the data before it enters the agent's context.
[COMMAND_EXECUTION]: The skill employs the Bash tool to perform project maintenance tasks such as updating session state files. This functionality, while appropriate for its technical director role, forms an attack surface when paired with the reading of external documentation.

create-architecture