tool-schema-design
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is primarily instructional, providing best practices for tool schema design to improve agent reliability and safety. No malicious patterns were identified across any analyzed files.
- [COMMAND_EXECUTION]: The skill includes a Python script
scripts/validate_tool_schema.pyintended for local use. Analysis of the script confirms it performs static analysis on JSON input using standard libraries (json,re,argparse). It does not execute shell commands or spawn subprocesses. - [CREDENTIALS_UNSAFE]: The skill actively promotes secret management best practices, instructing users to keep credentials server-side and out of model-facing schemas. The validation script includes heuristic checks to detect and warn about parameters that appear to be secrets (e.g.,
api_key,token,secret). - [EXTERNAL_DOWNLOADS]: No external network operations, package installations, or remote script fetches were detected. The skill operates entirely on local instructions and scripts.
- [PROMPT_INJECTION]: The instructions focus on hardening tool definitions against unintended model behavior. There are no attempts to override agent system prompts or bypass safety guardrails.
Audit Metadata