claude-agent-sdk
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides purely instructional documentation and code examples for the official Anthropic Claude Agent SDK (Python:
claude-agent-sdk, TypeScript:@anthropic-ai/claude-agent-sdk). All package references and installation instructions target official, trusted sources. - [SAFE]: No malicious patterns, obfuscation, or unauthorized access behaviors were detected. The content is technical documentation intended to assist developers in building agents.
- [SAFE]: The skill explicitly promotes security best practices, such as restricting tool sets to the minimum necessary requirements, warning against using unrestricted permission modes in production, and providing examples of hooks to block access to sensitive file paths like
.envor/etc. - [SAFE]: External network interactions described in the examples, such as sending notifications to Slack or fetching API data, are represented as standard developer functionality using generic placeholders and targeting well-known, trusted services.
Audit Metadata