@mulmoclaude/core@5.12.1
Shared server-side core for MulmoClaude and MulmoTerminal — the always-shipped-together subsystems consolidated behind subpath exports so the two hosts can't drift. Server-only except the browser-safe ./artifacts, ./whisper/client, ./workspace-setup/slug,
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe manifest has no install hook, and source review traced the flagged Google behavior to collection-scoped calendar synchronization and a Google OAuth flow. Static similarity signals did not correspond to a concrete attack in the inspected source.
Decision evidence
public snapshot- package.json defines build, typecheck, test, and lint scripts; it has no install lifecycle hook.
- The skill bridge mirrors allowlisted files from the workspace skill staging directory into that workspace’s Claude skills directory.
- Google calendar code keeps local sync state and pushes mapped collection records; it propagates deletions only when the collection enables that option.
- Google login uses the package’s configured HTTPS broker and checks that the returned authorization page is Google’s consent page.
- The flagged compiled files and inspected dependencies show module imports and ordinary utility, markdown, and Google authentication code; no concrete attack chain was found.
- Static similarity alerts identify matching source files but do not establish malicious behavior in the inspected source.
Source & flagged code
14 flagged · loading sourcePackage metadata claims a different repository identity while copied source loads a runtime dependency bridge.
dist/discovery-CbwHtcwk.jsView on unpkg · L1447Findings
Affected versions and remediation
This report applies to @mulmoclaude/core@5.12.1.
See all versions of @mulmoclaude/core for other recorded verdicts.
- Review the evidence and your use of @mulmoclaude/core@5.12.1 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.