art-bible
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface (Category 8).
- Ingestion points: The skill reads untrusted content from
design/gdd/game-concept.md(specifically the Visual Identity Anchor section) anddesign/art/art-bible.md. - Boundary markers: The skill interpolates extracted user data directly into instructions for sub-agents (e.g., in Phase 2, Section 1) without using protective delimiters or 'ignore embedded instructions' warnings.
- Capability inventory: The skill has the capability to write to the file system (
Write,Edit) and delegate tasks to other agents (Task), which could be influenced by malicious content in the ingested GDD files. - Sanitization: As a primary mitigation, the skill uses the
AskUserQuestiontool to present all agent-generated drafts to the user for explicit approval before any file-write operations occur.
Audit Metadata