integrate-arcjet-guard-cloudflare-think
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides legitimate integration instructions for Arcjet Guard into Cloudflare Think agents.
- [EXTERNAL_DOWNLOADS]: The skill references official documentation and examples from the vendor's website (arcjet.com) and GitHub organization (github.com/arcjet).
- [CREDENTIALS_UNSAFE]: The skill correctly advises using environment variables (
process.env.ARCJET_KEY) for API key management rather than hardcoding secrets. - [INDIRECT_PROMPT_INJECTION]: The skill implements a security boundary by screening inbound messages and tool calls.
- Ingestion points: User-provided messages (
userText) and tool arguments (input) are ingested inSKILL.md. - Boundary markers: Uses
arcjet.guard()to explicitly check content before processing. - Capability inventory: The skill controls the execution of agent tools via the
beforeToolCallhook. - Sanitization: Uses
policyInput.server.string()to wrap and validate external content.
Audit Metadata