workflow-greenfield-init
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by design due to its orchestration of data-gathering and processing tasks.
- Ingestion points: The workflow is initiated using the 'user's prompt as context' and explicitly calls the
$web-researchskill to perform deep-dive investigations into external, untrusted web sources. - Boundary markers: The skill body does not specify clear delimiters or instructions to ignore potential malicious content within the ingested research data or user prompts.
- Capability inventory: The skill possesses extensive capabilities, including spawning sub-agents (
spawn_agent), accessing internal configuration files (workflows.json,SKILL.mdfiles), and executing infrastructure-modifying skills like$scaffoldand$harness-setup. - Sanitization: There is no evidence of explicit sanitization or filtering of the external research data before it is passed to subsequent planning, architecture, and code-generation phases.
Audit Metadata