@mastra/server@1.61.0
⚠ Under reviewAI Security Review
scanned 6d ago · by lpm-firewall-aiThe inspected source contradicts the scanner’s specific credential-redirect claim, but does not cover the full flagged runtime surface or required paths. Since the unresolved review scope is material, the source evidence is insufficient for a clean or malicious verdict.
Decision evidence
public snapshot- The manifest points to built entrypoints and shows build and development scripts; the inspected excerpt does not establish an install hook.
- The gateway provider defaults to the Vercel AI Gateway and obtains its gateway authentication token for requests.
- The inspected Anthropic token-estimation helper uses the Anthropic key and sends its request to Anthropic’s API.
- A package-install helper spawns a package manager when its install operation is called.
- The inspected excerpts do not substantiate the scanner’s claim that an Anthropic credential is sent to an unrelated gateway or that a redirect is persisted.
- The runtime review remains incomplete: the flagged bundles and required runtime paths were not all traced.
Source & flagged code
6 flagged · loading sourcePackage source references child process execution.
dist/dist-DERatX43.cjsView on unpkg · L30Package source references a known benign dynamic code generation pattern.
dist/schema-Bv9OdW1f-CxTiW8RD.cjsView on unpkg · L3778Findings
Affected versions and remediation
This report applies to @mastra/server@1.61.0.
See all versions of @mastra/server for other recorded verdicts.
- Review the evidence and your use of @mastra/server@1.61.0 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.