@zaip-ai/cli@1.3.4
CLI oficial para instalar e configurar o ZAIP-AIOS em qualquer repositório
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThis is an explicit-user-command AI-agent configuration mutator with unusually broad defaults, not an install-time malware chain. Its capabilities warrant a warning, but the reviewed source does not establish malicious behavior.
Decision evidence
public snapshot- The explicit init command installs a project template and normally enables global Codex configuration changes.
- Generated project Codex configuration sets a full-access profile as the default.
- That profile grants root filesystem writes, unrestricted networking, inherited environment variables, and enabled browser/computer-use plugins.
- The installer writes Codex configuration and hook files into the target project.
- The global configuration reconciler adds a PreToolUse hook to the user's Codex config.
- package.json has no preinstall, install, or postinstall lifecycle hook.
- All configuration mutation is reached through explicit CLI commands, not package installation.
- The reviewed hook source is a command safety filter; no credential exfiltration or remote payload download was found.
- The shipped Windows executable is accompanied by source that launches a local design runtime; the design server binds to loopback only.
Source & flagged code
10 flaggedPackage hides binary, compressed, or executable-looking payloads in test/fixture/hidden paths.
templates/.agents/design/zaip-design.exeView on unpkgPackage ships native binary artifacts.
templates/.agents/design/zaip-design.exeView on unpkgFindings
Affected versions and remediation
This report applies to @zaip-ai/cli@1.3.4.
See all versions of @zaip-ai/cli for other recorded verdicts.
- Review the evidence and your use of @zaip-ai/cli@1.3.4 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.