browser-harness-authoring
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust safety framework for browser automation. It explicitly instructs the agent to treat all website content as untrusted and provides a standalone validator script to ensure generated output remains free of credentials and private information.
- [COMMAND_EXECUTION]: The skill uses a bundled Python script,
validate_harness.py, to perform structural and safety checks on generated documents. This is a local, utility-based execution intended for quality assurance and verification. - [DATA_EXPOSURE]: While the validator script contains regular expressions for identifying sensitive patterns (such as API keys and private keys), these are utilized solely for scanning purposes to prevent data leakage in authored harnesses rather than representing hardcoded secrets.
Audit Metadata