anthropic-agent-sdk
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of documentation files, tutorials, and code examples for building agents using the Claude Agent SDK. No executable code is distributed with the skill itself.
- [DATA_EXPOSURE]: References to sensitive file paths (e.g.,
.env,~/.aws/credentials,.ssh/id_rsa) and credentials appear strictly within security hardening guides and tutorials. These examples are used to teach developers how to protect sensitive data and prevent unauthorized access. - [COMMAND_EXECUTION]: The shell commands included in the documentation (such as
npm install,pip install, or Docker execution strings) are standard procedures for installing and deploying the official Anthropic SDK and its dependencies. - [REMOTE_CODE_EXECUTION]: The documentation mentions the official Anthropic installation script (
https://claude.ai/install.ps1). As this is a well-known service and the official source for the tool, this reference is documented neutrally as a legitimate installation method. - [DYNAMIC_CONTEXT_INJECTION]: While the skill documents the
!commandsyntax used by the underlying platform, it does not employ the syntax for any actual execution within the skill's own instructions. Documentation of this feature is intended for developer education.
Audit Metadata