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
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:57 AM
Security Audit — agent-trust-hub — runaway-guard