cost-aware-llm-pipeline
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a pattern library for cost-effective LLM application development, using standard libraries like Anthropic for API interaction.
- [SAFE]: No hardcoded credentials, sensitive file access, or malicious external connections were found in the instructions or code snippets.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a pipeline for processing external data, which is an inherent attack surface for LLM applications. 1. Ingestion points: The
textparameter in theprocessfunction withinSKILL.mdreceives external data. 2. Boundary markers: Not explicitly implemented in the example code, although prompt caching structures are used. 3. Capability inventory: The pipeline performs API calls to LLM providers using theanthropicclient. 4. Sanitization: Not featured in the cost-optimization patterns provided.
Audit Metadata