@minimax-ai/code@0.2.4
Minimax Code — terminal coding agent
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThe only install-time hook is a local dependency compatibility check, not a persistence or control-surface mutation. The inspected helper behavior is user-invoked tooling rather than an automatic attack path.
Decision evidence
public snapshot- The package declares the mcode command-line entrypoint and identifies itself as a terminal coding agent.
- Its postinstall hook runs only the bundled native-install verifier.
- The required PDF helper launches Playwright and only attempts a Chromium install after its own launch fails.
- No credential harvesting, unrelated exfiltration, destructive action, or foreign agent-control mutation was identified in the inspected lifecycle code.
- The large CLI bundle contains product networking and agent functionality, but no concrete malicious path was established by static inspection.
Source & flagged code
11 flaggedPackage defines install-time lifecycle scripts.
package.jsonView on unpkgPackage source references child process execution.
assets/skills/pdf/scripts/render_html.cjsView on unpkg · L95Package source references shell execution.
assets/skills/pdf/scripts/render_html.cjsView on unpkg · L127Findings
Affected versions and remediation
This report applies to @minimax-ai/code@0.2.4.
See all versions of @minimax-ai/code for other recorded verdicts.
- Review the evidence and your use of @minimax-ai/code@0.2.4 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.