gsp-screenshot-critic
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on external visual input and includes instructions to modify local project files, creating a theoretical vector for indirect prompt injection.
- Ingestion points: Processes screenshots and visual captures provided by the user (SKILL.md).
- Boundary markers: No explicit delimiters or instructions are provided to the agent to differentiate between visual content and instructional context, nor to ignore instructions embedded within the images.
- Capability inventory: The skill possesses the capability to write or append critique findings to
docs/game-studio/quality-report.md(SKILL.md). - Sanitization: There are no explicit instructions for the agent to sanitize, validate, or escape the content derived from the screenshots before writing it to the local file system.
Audit Metadata