claude-api
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation references official resources from trusted organizations, specifically Anthropic's own domains (platform.claude.com) and GitHub repositories (github.com/anthropics). These are well-known, established services for the tool's context.
- [COMMAND_EXECUTION]: The skill contains standard developer instructions for installing tools (e.g., using 'sudo tar' for CLI installation) and utilizing the 'bash' tool for coding tasks. These commands are typical for the stated purpose of assisting with API integration and are not used maliciously.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to interpret user prompts and provide API-related guidance. While this involves an inherent attack surface for indirect injection (as the agent reads documentation and potentially user-supplied code), the instructions include safeguards such as scanning for non-Anthropic markers and emphasizing the use of official SDKs.
Audit Metadata