runaway-guard
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves exclusively as a documentation and policy framework for 'Cost-safety discipline' (FinOps). It does not contain executable code, scripts, or automated triggers that perform network operations or file system modifications.
- [SAFE]: References to external AI SDKs (such as OpenAI, Anthropic, and Fal.ai) and platforms (like Inngest) are used for contextual examples and to define the scope of the safety guidelines.
- [SAFE]: The provided code snippets are illustrative examples of safe vs. unsafe coding patterns. They are not intended for direct execution by the agent platform but rather as templates for the agent to follow when assisting a developer.
- [SAFE]: No obfuscation, data exfiltration patterns, or malicious prompt injection techniques were detected. The skill's metadata and author information are consistent with its stated purpose.
Audit Metadata