edge-computing-patterns
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS] (MEDIUM): The skill references several remote tools and packages from organizations not included in the [TRUST-SCOPE-RULE] whitelist.
- Evidence: References to
npm create cloudflare@latest,npx wrangler, and Deno Deploy patterns inSKILL.mdandreferences/cloudflare-workers.md. - Context: While these are standard industry tools, they are classified as unverifiable dependencies because
cloudflareanddenoare not explicitly listed as trusted organizations. - [EXTERNAL_DOWNLOADS] (LOW): The skill references Vercel Edge functions and Vercel-specific patterns.
- Context: Vercel is a trusted organization per the [TRUST-SCOPE-RULE], so these references are downgraded to LOW/INFO.
- [CREDENTIALS_UNSAFE] (SAFE): The skill contains no hardcoded secrets or keys. It explicitly advises users to 'Use environment variables for secrets (never hardcode)' and correctly demonstrates the use of secret management tools like
wrangler secret put. - [PROMPT_INJECTION] (SAFE): No malicious instructions designed to override agent behavior or bypass safety filters were detected in the instructions or metadata.
- [DATA_EXFILTRATION] (SAFE): No unauthorized data transfer or exfiltration patterns were found. Network operations (fetch) are standard for the edge runtime context described.
Audit Metadata