cost-aware-llm-pipeline

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill provides educational code snippets for cost management using official Anthropic SDK patterns.
  • [INDIRECT_PROMPT_INJECTION]: While the skill defines functions that process external text inputs (e.g., the process and build_cached_conversation functions), these are provided as implementation templates for cost routing and do not include instructions for the agent to execute untrusted commands. Standard LLM safety guardrails apply at the API level.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 10:29 AM
Security Audit — agent-trust-hub — cost-aware-llm-pipeline