@oranlens/oranleads
橙客 OranLeads:社媒评论区 AI 获客工作台(DSH 插件)
This report shows the evidence for @oranlens/oranleads@0.8.4. 0 of 7 known versions are blocked or flagged. See all versions.
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe reviewed network and credential flows fit the documented lead-research plugin: it reports disclosed anonymous business-profile statistics and downloads its browser assets from documented hosts. The package has no install-time hook, and I found no evidence of unrelated credential transfer or unrequested account actions.
Decision evidence
public snapshot- The package is a DSH social-media lead research plugin; its host and client are the declared entry points.
- The only lifecycle hook shown is prepublishOnly, not an install-time hook.
- The README discloses anonymous reporting of five business-profile fields to the shop backend.
- The report POST omits credentials and sends a profile payload; its endpoint can be overridden through OL_BUSINESS_REPORT_URL.
- The browser download URLs are identified as the CloakBrowser CDN and its OSS fallback.
- The host submits a business-profile report during startup, so this is runtime telemetry; the README describes it as anonymous statistics.
- The package uses platform cookies for its stated browser-based collection workflow; I found no source evidence that it forwards them to an unrelated host.
Source & flagged code
3 flaggedPackage source references dynamic require/import behavior.
host/index.jsView on unpkg · L1Package ships native binary artifacts.
native/bin/win32-arm64/oranleads-core.nodeView on unpkgFindings
Affected versions and remediation
This report applies to @oranlens/oranleads@0.8.4.
The versions list below shows the verdict for each recorded version.
- Review the evidence and your use of @oranlens/oranleads@0.8.4 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.