schema-pruned-function-calling
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of documentation and a Python helper function designed to optimize JSON schemas for tool-calling agents by removing decorative metadata and verbose descriptions.
- [DYNAMIC_EXECUTION]: The Python code provided in the skill is a static utility for data transformation. It does not perform dynamic code loading, compilation, or execution of untrusted input. It exclusively uses standard dictionary methods to filter and restructure data.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a protocol for processing external tool schemas. While this involves handling data that may originate from third-party sources (MCP servers or tool registries), the provided implementation acts as a sanitization layer that reduces the amount of natural language passed to the model, thereby decreasing the potential surface area for injection attacks within tool descriptions.
- [DATA_EXFILTRATION]: No network operations, sensitive file path access, or credential harvesting patterns were detected. The logic is entirely local to the provided function context.
Audit Metadata