@sciagent/cli@1.0.62
SciAgent CLI - AI Research Assistant
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThe package has an unconsented postinstall payload-provisioning chain with weak artifact verification and global dependency mutation. Source inspection does not establish that the retrieved payload is malicious or that a foreign AI-agent control surface is hijacked, so blocking is not justified.
Decision evidence
public snapshot- `package.json` invokes `scripts/postinstall.js` automatically.
- `scripts/postinstall.js` downloads native CLI and CodeBuddy binaries during postinstall.
- Downloads are written executable under `~/.sciagent/bin` or `%LOCALAPPDATA%\sciagent\bin`.
- Remote artifacts are accepted by size only; no hash or signature check is present.
- Postinstall runs `npm install -g @tencent-ai/codebuddy-code` as a fallback.
- `bin/sciagent.js` later spawns the downloaded local binary.
- No credential, environment-value, SSH, or browser-data harvesting found.
- No outbound upload or exfiltration logic found.
- No destructive file operations beyond temporary-download cleanup found.
- No writes to Claude, Cursor, Copilot, MCP, or other agent control-surface configuration found.
- The launcher executes a SciAgent path only after an explicit `sciagent` command.
Source & flagged code
5 flagged · loading sourcePackage defines install-time lifecycle scripts.
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 @sciagent/cli@1.0.62.
See all versions of @sciagent/cli for other recorded verdicts.
- Review the evidence and your use of @sciagent/cli@1.0.62 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.