unity-game-qa-performance
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection by processing external data and documents to drive tool execution.
- Ingestion points: The instructions direct the agent to read various untrusted external sources, including "Work Item", "验收" (acceptance criteria), "GDD/TDD", "数值" (numerical values), "资源" (resources), and "音频与发行文档" (audio and issuance documents) (SKILL.md).
- Boundary markers: There are no specific delimiters, framing techniques, or instructions for the agent to ignore embedded commands within these external data sources.
- Capability inventory: The agent is granted the capability to perform builds and run tests using the
unity:unity-clitool based on the ingested data (SKILL.md). - Sanitization: The skill lacks any instructions for the validation, escaping, or sanitization of content retrieved from external documents before it is processed or used to drive tool interactions.
Audit Metadata