infrastructure-claude-code-game-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references several external tools and MCP servers that are fetched via
npxandbunx. These include the GitHub MCP server and Upstash Context7, as well as standard development utilities like Biome and Drizzle ORM. - [COMMAND_EXECUTION]: Includes templates for shell-based hooks that automate routine tasks like code formatting and provide keyword-based routing to relevant documentation when specific domains (e.g., narrative, database) are detected in prompts.
- [DYNAMIC_EXECUTION]: Utilizes the agent platform's hook system to execute shell logic at specific triggers. This is used defensively to implement guards that detect and block potentially harmful database commands like 'DROP TABLE' or 'TRUNCATE'.
- [INDIRECT_PROMPT_INJECTION]: Employs prompt and tool input inspection within hooks to enhance the developer experience. This surface is used for safety filtering and guiding the agent to relevant project skills.
Audit Metadata