weave
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted business requirements and specifications to generate design artifacts like state machine definitions and Saga compensation tables.
- Ingestion points: Requirements enter via
USER_TO_WEAVE_HANDOFFand specification sections fromSCRIBE_TO_WEAVE_HANDOFF(e.g., inreference/handoffs.md). - Boundary markers: The skill does not define specific delimiters or instructions to prevent user input from overriding agent logic.
- Capability inventory: The skill generates complex workflow logic that is passed to other agents (e.g.,
Builder) for code implementation, potentially allowing a malicious requirement to influence generated code. - Sanitization: There is no evidence of sanitization or strict schema validation for the natural language requirements processed.
Audit Metadata