ai-coding-agents-settings-policy
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill contains references to documentation and official repositories from trusted organizations, including Anthropic (code.claude.com) and OpenAI (openai.com). These links are provided for technical reference and architectural comparison and do not involve the execution of untrusted remote code.
- [COMMAND_EXECUTION]: The files include various code snippets (Python, TOML, JSON) illustrating implementation logic for settings merging, trust classes, and validation schemas. These snippets are non-executable documentation templates and pose no runtime risk.
- [SAFE]: The primary focus of the skill is to provide defensive security guidance, such as implementing 'managed policy' constraints that cannot be bypassed by user-level configuration and enforcing trust-class restrictions on plugin surfaces to prevent unauthorized capability expansion.
Audit Metadata