@mastra/server@1.77.0
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe source exposes agent-invoked shell execution, a powerful dual-use capability that warrants a warning. The flagged credential handling is consistent with authentication to the Vercel AI Gateway and does not establish credential theft.
Decision evidence
public snapshot- The agent builder exposes an execute-command tool that accepts a command and working directory.
- The tool passes the supplied command to child_process.exec and can return its output.
- The flagged credential path uses the Vercel AI Gateway provider, with a configurable base URL and a gateway-specific token.
- The top-level package entrypoint throws and directs callers to specific subpaths.
Source & flagged code
6 flagged · loading sourceThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
dist/dist-Cfv3ewR4.cjsView on unpkgPackage source references child process execution.
dist/dist-Cfv3ewR4.cjsView on unpkg · L30Findings
Affected versions and remediation
This report applies to @mastra/server@1.77.0.
See all versions of @mastra/server for other recorded verdicts.
- Review the evidence and your use of @mastra/server@1.77.0 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.