agents-chat-cli@3.39.0
多智能体群聊工具 - 支持 OpenCode/Claude Code/Codex/pi 内核,微信风格聊天界面
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThe package contains a network-reachable, unauthenticated shell endpoint and an automatic global install lifecycle action. The code supports a critical remote-command-execution risk, but the inspected source does not establish covert malware intent.
Decision evidence
public snapshot- The manifest runs a postinstall hook automatically.
- The postinstall hook calls kernel setup, which globally installs opencode-ai when no supported AI CLI is found.
- The HTTP terminal endpoint writes request data directly to a spawned shell's standard input without authentication.
- The server listens without a loopback host restriction, so the terminal endpoint can be reachable from the network.
- The service can also write OpenCode's global configuration, including a custom provider endpoint and key.
- No package self-dependency or staged payload was found.
- Registry update checks target this package's npm registry endpoint.
- The AI CLI execution and configuration features are consistent with the package's stated chat-agent function, although their exposure is unsafe.
Source & flagged code
19 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgA bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to agents-chat-cli@3.39.0.
See all versions of agents-chat-cli for other recorded verdicts.
- Review the evidence and your use of agents-chat-cli@3.39.0 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.