@minimax-ai/code
Minimax Code — terminal coding agent
This report shows the evidence for @minimax-ai/code@0.7.4. 0 of 56 known versions are blocked or flagged. See all versions.
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe TUI sends runtime telemetry by default, including a persistent device ID and user or account identifiers alongside launch metadata. The inspected payload contains no prompts, file contents, or credentials, so this warrants a warning for undisclosed telemetry rather than a publish block.
Decision evidence
public snapshot- The TUI initializes business telemetry by default and queues report events for POST to a region- and build-specific reporting endpoint.
- Telemetry includes a MiniMax user ID when available, a service op-group ID, a persistent random device ID, platform, app version, and event properties.
- The TUI records a launch event with launch type and user ID; the README does not disclose this telemetry or describe a way to turn it off.
- The documented package purpose is an interactive coding agent that edits code and runs commands.
- The postinstall check runs a SQLite smoke query against an in-memory database.
- The telemetry fields inspected are usage and device metadata; no prompts, file contents, or credentials were present in the inspected event construction.
- Command execution is part of the package’s documented coding-agent purpose.
- The postinstall script’s inspected SQLite check is limited to an in-memory query.
Source & flagged code
24 flaggedPackage defines install-time lifecycle scripts.
package.jsonView on unpkgPackage source references child process execution.
chunks/chunk-L7HIW3OD.jsView on unpkg · L1Findings
Affected versions and remediation
This report applies to @minimax-ai/code@0.7.4.
The versions list below shows the verdict for each recorded version.
- Review the evidence and your use of @minimax-ai/code@0.7.4 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.