The Agent Skills Directory

[SAFE]: The skill primarily uses Read, Glob, and Grep to analyze existing game design documents within the project's local directory structure.\n- [SAFE]: It implements a multi-agent review process using the Task tool to spawn specialist subagents. This delegation is used for logical analysis of document content rather than execution of remote or untrusted scripts.\n- [SAFE]: Write operations to files such as systems-index.md or review logs are protected by the AskUserQuestion tool, ensuring the user must explicitly approve any modifications to the project state.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external design documents which could potentially contain malicious instructions.\n
Ingestion points: Target design document and referenced GDDs read via Read and Glob tools in Phase 1.\n
Boundary markers: Absent; content is passed directly to the model and subagents for review.\n
Capability inventory: Write, Edit, and Task tools.\n
Sanitization: None; the skill relies on the core model's ability to distinguish between data (the GDD) and instructions (the review task).

design-review