write-flow

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill implements a structured workflow for scaffolding Windmill flows using the official wmill CLI, emphasizing the use of the flow new command over manual file creation to ensure consistency.\n- [SAFE]: Explicit safety instructions are provided to prevent the agent from performing actions with side effects (like wmill sync push or opening visual previews) without explicit user approval or request.\n- [SAFE]: The skill includes a complete OpenFlow JSON schema, which serves as a strict technical constraint to ensure the agent generates valid, well-structured flow definitions.\n- [COMMAND_EXECUTION]: The skill utilizes the wmill CLI for flow management, which is the intended primary purpose of the skill. The instructions show proper usage of the CLI for local development and iteration, such as creating new flows, previewing them locally, and generating metadata.\n- [SAFE]: No evidence of credential theft, data exfiltration, or persistence mechanisms was detected. The instructions for handling resources ($res:) and environment variables follow standard Windmill security practices and encourage user oversight for deployments.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 08:10 AM
Security Audit — agent-trust-hub — write-flow