prompt-caching
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a technical guide for optimizing LLM performance and cost. The included TypeScript/JavaScript examples follow standard practices for caching and API interaction.
- [EXTERNAL_DOWNLOADS]: The skill references standard, well-known libraries including
@anthropic-ai/sdkfor interacting with Claude models andioredisfor Redis connectivity. These are established tools in the ecosystem. - [CREDENTIALS_UNSAFE]: No hardcoded secrets or API keys were found. The code correctly demonstrates accessing sensitive configuration like
REDIS_URLthrough environment variables. - [COMMAND_EXECUTION]: The skill does not contain any shell commands, subprocess executions, or dynamic code evaluation (eval/exec) patterns.
- [DATA_EXFILTRATION]: There is no evidence of unauthorized data access or transmission to unknown third-party domains. Network operations are restricted to well-known service endpoints (Anthropic API and user-configured Redis instances).
Audit Metadata