arkiv
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides clear security protocols for the agent, such as never requesting private keys and ensuring RPC access keys are kept in server-side headers rather than prompts or URLs.
- [SAFE]: It includes defensive instructions against indirect prompt injection, specifically directing the agent to treat external database payloads and attributes as untrusted data.
- [SAFE]: All external links and package references (such as @arkiv-network/sdk and viem) are to legitimate, well-known, or vendor-official sources. No malicious code execution or exfiltration patterns were detected.
- [SAFE]: The skill uses a routing architecture to delegate tasks to specialized sub-skills, promoting a modular and manageable security posture.
Audit Metadata