runaway-guard
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional and serves as a best-practice guide for cost management in AI-integrated applications.
- [SAFE]: All referenced external services (such as OpenAI, Anthropic, Fal.ai, and Replicate) are well-known technology providers. The skill directs users to official dashboards for billing configuration.
- [SAFE]: No malicious patterns such as prompt injection, credential exfiltration, or unauthorized command execution were detected.
- [SAFE]: Code snippets are illustrative examples of safe programming patterns, including the use of idempotency keys and bounded loops to prevent infinite billing cycles.
Audit Metadata