registry / @vercel/cli-support / 65.4.0
@vercel/cli-support@65.4.0
⚠ Under reviewThe JavaScript Vercel CLI, installed by `vercel` where no native binary is available
AI Security Review
scanned 9d ago · by lpm-firewall-aiThe inspected source excerpts do not substantiate the scanner’s host-fingerprint or credential-exfiltration claims. Required runtime callers and the complete citation coverage set remain unresolved, so the package cannot be cleared or blocked on this inspection.
Static reason
One or more suspicious static signals were detected.
Trigger
No malicious trigger was verified in the inspected excerpts.
Impact
No concrete harmful impact was established from the inspected excerpts.
Mechanism
The inspected code includes Vercel OIDC authentication and CLI configuration behavior; these excerpts do not establish a harmful data flow.
AI rationale
The inspected source excerpts do not substantiate the scanner’s host-fingerprint or credential-exfiltration claims. Required runtime callers and the complete citation coverage set remain unresolved, so the package cannot be cleared or blocked on this inspection.
Evidence
package.jsondist/chunks/chunk-UVHKXCR5.jsdist/commands-ai.jsdist/chunks/chunk-PFPJLIFF.jsdist/commands-core.js
Network endpoints1
api.vercel.com/v1/projects/${projectId}/token
Decision evidence
public snapshotAI called this Manual Review at 62.0% confidence as Unknown with medium false-positive risk.
AI review observations
- The manifest identifies the package and version; inspected metadata shows no lifecycle hooks in the excerpted portion.
- The inspected OIDC helper sends an authentication token to the Vercel API for an OIDC token refresh.
- The inspected agent-related excerpts read Claude’s installed-plugin registry and update Vercel best-practices content.
Evidence against
- The cited excerpts do not establish host-identity collection or credential transfer to an unrelated recipient.
- The inspected code does not establish unconsented agent configuration changes.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShellWebSocket
HighEntropyStringsMinifiedPossibleObfuscationTelemetryUrlStrings
Source & flagged code
12 flaggeddist/commands-blob.jsView file
8patternName = private_key_rsa
severity = critical
line = 8
Critical
Critical Secret
Package contains a critical-looking secret pattern.
dist/commands-blob.jsView on unpkg · L88patternName = private_key_rsa
severity = critical
line = 8
Critical
dist/commands-ai.jsView file
55`}}],envExports:[{name:GATEWAY_API_KEY_ENV,value:ctx.apiKey}],notes:[`Mistral Vibe will default to ${activeModel} through the gateway; switch with /model.`,...models3.length?[`${mo...
L56: ${import_chalk19.default.dim(" Detected agents are pre-selected \xB7")}${CHECKBOX_INSTRUCTIONS}`,instructions:!1,loop:!1,pageSize:Math.max(choices.length,7),choices})).map(id=>get...
L57: `)}function preservedEnvLines(current,writtenNames){if(!current)return[];let start=current.indexOf(MANAGED_BLOCK_START),end=current.indexOf(MANAGED_BLOCK_END);if(start===-1||end===...
Findings
2 Critical8 High4 Medium6 Low
CriticalCritical Secretdist/commands-blob.js
CriticalSecret Patterndist/commands-blob.js
HighChild Processdist/commands-ai.js
HighShell
HighSame File Env Network Executiondist/commands-ai.js
HighHost Fingerprint Exfiltrationdist/chunks/chunk-UVHKXCR5.js
HighCommand Output Exfiltrationdist/commands-ai.js
HighSandbox Evasion Gated Capabilitydist/commands-core.js
HighCross File Remote Execution Contextdist/chunks/chunk-6YXIWZ3X.js
HighSemantic Analysis Limiteddist/chunks/chunk-4NTBYNFR.js#virtual:normalized:round1
MediumDynamic Requiredist/get-latest-worker.cjs
Medium
Affected versions and remediation
This report applies to @vercel/cli-support@65.4.0.
See all versions of @vercel/cli-support for other recorded verdicts.
- Review the evidence and your use of @vercel/cli-support@65.4.0 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.