soia-env-kimi-cli-install

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches version metadata and release information from well-known and official services, including registry.npmjs.org, pypi.org, and the GitHub API (api.github.com).
  • [COMMAND_EXECUTION]: Local scripts execute standard system commands such as shutil.which to locate executables and subprocess.run to call kimi --version or npm prefix -g. These are used exclusively for environment inspection and installation path discovery.
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions for installing the @moonshot-ai/kimi-code package. It implements a security-first approach by instructing the agent to download remote installation scripts to a temporary directory and verify their content before local execution, rather than piping network responses directly into a shell.
  • [SAFE]: The skill includes multiple proactive safety constraints: it prioritizes user-level installations to avoid sudo, implements path sanitization to prevent directory traversal when displaying status, and strictly prohibits the agent from reading or printing sensitive tokens or credentials stored in the config.toml file.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 08:46 AM
Security Audit — agent-trust-hub — soia-env-kimi-cli-install