converge
Warn
Audited by Snyk on Jul 13, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.75). The required workflow is the Go command
ao converge, which (per SKILL.md) dispatches judge legs and reads “the judges' reasons” between rounds; those reasons are outsider-authored free text produced by external judge models/panes at runtime and then fed back into the agent’s LLM context.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata