clade-security-basics
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill body contains phrases associated with prompt injection, such as "Ignore your instructions" and "Reveal your system prompt". These are used exclusively as examples of adversarial inputs that developers should program their systems to reject, following best practices for LLM security.- [CREDENTIALS_UNSAFE]: The skill references API key placeholders (e.g., "sk-ant-...") within code blocks. These are used to illustrate the difference between insecure hardcoding and secure environment variable storage; no actual credentials are leaked.- [EXTERNAL_DOWNLOADS]: The instructions recommend the use of the "@upstash/ratelimit" library for implementing rate limits. Upstash is a well-known service provider, and the reference serves as a standard example of architectural defense against resource exhaustion attacks.
Audit Metadata