claude-cookbooks
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The contribution guide in
references/CONTRIBUTING.mdprovides a standard installation command for theuvpackage manager usingcurlpiped to a shell. The script is hosted athttps://astral.sh/uv/install.sh, which is the official distribution point for the well-known Astral ecosystem tool. - [EXTERNAL_DOWNLOADS]: The skill documentation describes the installation of several well-known Python packages for development and testing, including
anthropic,ruff,pre-commit,jupyter, andnbconvertvia official package registries. - [COMMAND_EXECUTION]: The skill includes documentation for using standard CLI tools such as
gitfor repository management,ghfor GitHub interactions, and package managers likepipanduvfor environment setup. - [INDIRECT_PROMPT_INJECTION]: The
scripts/memory_tool.pyutility allows an agent to read, write, and manage files in a local storage directory. This capability presents a surface where external or untrusted data processed by the agent could be stored and subsequently re-read into the agent's context. The implementation mitigates this risk by enforcing strict path validation usingpathlibto prevent directory traversal and restricting file operations to a specific/memoriesdirectory.
Audit Metadata