gameg
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-supplied input through the
$ARGUMENTSvariable without implementing boundary markers (such as XML tags or unique delimiters) or explicit instructions to treat the content as data rather than instructions. This creates a surface where a user could potentially provide input designed to override the agent's logic, although the skill's restricted scope as a dispatcher limits the potential impact. - Ingestion points:
$ARGUMENTSparameter inSKILL.md. - Boundary markers: Not implemented.
- Capability inventory: The skill directs the agent to load and follow another skill file (
../game-generalist/SKILL.md) and read a reference file (references/intent-brief.md). - Sanitization: No sanitization or escaping of the user-provided idea is performed before it is integrated into the prompt instructions.
Audit Metadata