3dg
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided input through the
$ARGUMENTSvariable and interpolates it directly into the instruction flow without sanitization or boundary markers. - Ingestion points: The user's message is captured via the
$ARGUMENTSplaceholder inSKILL.mdand included in the text processed by the agent. - Boundary markers: No delimiters, such as XML tags or clear separators, are used to isolate the untrusted user input from the skill's logic.
- Capability inventory: The skill instructs the agent to load and follow another skill file (
../3d-generalist/SKILL.md) and run an intent brief fromreferences/intent-brief.md. While this specific skill does not define dangerous tools, it passes potentially malicious instructions to the next stage of the agent's workflow. - Sanitization: There is no logic to filter or escape instructions provided within the user arguments, allowing a user to potentially attempt to override the agent's behavior during the transition to the 3D generalist skill.
Audit Metadata