coloring-page
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the assistant to execute a local binary
bin/coloring-pagewith user-supplied file paths, which is the intended core functionality of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-uploaded images, creating an attack surface where adversarial content in the image could attempt to influence the agent's behavior.
- Ingestion points: User-provided image files (
.jpg,.png,.webp) passed via the--inflag in SKILL.md. - Boundary markers: No explicit boundary markers or instructions to ignore embedded content are defined in the skill instructions.
- Capability inventory: The skill utilizes the
bin/coloring-pagecommand to process files (SKILL.md). - Sanitization: There is no explicit description of file path or image metadata sanitization in the skill metadata.
Audit Metadata