workflow-builder
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: This skill ingests untrusted user requirements to generate runnable workflow scripts, which constitutes an indirect prompt injection attack surface.
- Ingestion points: User instructions defining the workflow adapter logic (SKILL.md).
- Boundary markers: The skill instructions advocate for "explicit inputs and outputs" and "at-most-once dispatch" as safety measures (SKILL.md).
- Capability inventory: The skill is intended to produce a
workflow-script, involving code generation and potential file system effects (SKILL.md). - Sanitization: The guidelines suggest including a "dry-run or fixture demonstrating exact dispatch count" to allow for validation of the generated logic (SKILL.md).
Audit Metadata