game-code-review
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external repositories, which is a common surface for indirect prompt injection.
- Ingestion points: The skill is instructed to read repository instructions, diffs, task acceptance criteria, and design documents (SKILL.md, Phase 1).
- Boundary markers: No specific delimiters or instructions to ignore instructions embedded within the ingested code or documents are provided.
- Capability inventory: The skill uses a custom command syntax
$task-bugto report findings to a task management system (SKILL.md, Phase 3). - Sanitization: There are no instructions for sanitizing or escaping the content read from the repository before it is processed by the agent.
Audit Metadata