alloy-composition-search
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external, untrusted sources including patents and research papers via MCP tools like
document_to_alloy. This ingestion surface could potentially allow malicious instructions embedded in documents to influence the agent's behavior. - Ingestion points: Processes external document content retrieved via the
mace-mcpserver tools (SKILL.md). - Boundary markers: The skill instructions do not explicitly define delimiters or warnings to ignore instructions found within retrieved document data.
- Capability inventory: The skill is restricted to data retrieval and formatting; it does not utilize high-risk capabilities such as shell access, file system modifications, or arbitrary network operations.
- Sanitization: No explicit sanitization or validation of the retrieved document content is described in the skill instructions.
Audit Metadata