skills/connorads/dotfiles/typescript/Gen Agent Trust Hub

typescript

Pass

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [SAFE]: The skill exclusively contains educational documentation and engineering idioms for TypeScript development, focusing on architectural patterns and type safety.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines clear patterns for handling untrusted data, specifically the 'parse, don't validate' idiom. It recommends using schema libraries (such as Zod, Valibot, or Effect Schema) to refine external inputs into domain types at the application boundary, which serves as a defensive measure against malformed or malicious data injection.
  • [DYNAMIC_EXECUTION]: The instructions explicitly advise against top-level side effects in modules, such as opening network connections or reading environment variables at import time. This practice limits the risk of hidden or unexpected execution when modules are loaded into the agent's environment.
  • [SAFE]: The skill promotes the use of a Redacted<T> wrapper for sensitive information (tokens, keys, passwords), ensuring that credentials are only unwrapped at the point of use and not accidentally exposed in logs or telemetry.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 16, 2026, 05:06 PM