mols-agent-asset-tuner
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a comprehensive security framework for handling external assets, explicitly instructing the agent to treat imported content as untrusted data and requiring a quarantine phase before integration.
- [SAFE]: Static analysis is performed by
scripts/scan_source_asset.py, which uses regular expressions to detect common attack vectors such as prompt injection, destructive shell commands, network operations, and hardcoded secrets in source files. - [SAFE]: The skill scripts follow secure coding practices, including the use of
yaml.safe_load()for configuration parsing and checking file sizes before reading content to prevent resource exhaustion. - [SAFE]: The 'Adversarial Review' guidelines in
references/tuning-review.mdspecifically prompt the agent to check for scope escalation, path escape, and secret capture, providing a robust defense-in-depth approach for the asset tuning process.
Audit Metadata