Claude Zeroclaw
Security checks across malware telemetry and agentic risk
Overview
The package largely does what it claims (a local Claude Code daemon), but there are multiple inconsistencies and risky elements—an opaque/obfuscated setup script, a remote Windows MSI URL, missing declared dependencies, and memory heuristics that may capture secrets—that merit caution before installing or running it.
Do not run the one-line Windows command that calls msiexec on https://cloudcraftshub.com/api. Before installing or running this skill: - Inspect the code locally (especially scripts/setup.js which is heavily obfuscated). Avoid running obfuscated installers. - Prefer cloning the GitHub repository and reviewing scripts/setup.js and install.sh. If you must run installers, run them from trusted release artifacts (GitHub releases, signed packages). - Check package.json and the actual dependencies; run npm install only after you confirm what will be installed. - Be aware the daemon persists memories and run history in ~/.zeroclaw-claude and has heuristics that can elevate and store lines containing 'password', 'token', 'key' — avoid sending secrets through this system or disable memory collection. - During initial configuration, choose the most restrictive security level (readonly or standard) and avoid granting shell execution or 'full' mode unless you fully audit the code and trust it. - Consider running the installer and daemon inside an isolated environment (container or VM) until you are comfortable with behavior. - If you need help auditing specific files (the obfuscated scripts/setup.js, the Windows MSI endpoint, or memory extraction logic), get those inspected by a developer or security reviewer before proceeding.
SkillSpector
SkillSpector findings are pending for this release.
VirusTotal
No VirusTotal findings