@mulmoclaude/core@4.8.0
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 1mo ago · by lpm-firewall-aiThis is not malicious by static inspection, but it exposes a guarded mechanism for modifying an AI-agent control surface. Because the mechanism requires an explicit host call and is path- and content-type-restricted, blocking publication is not warranted.
Decision evidence
public snapshot- The exported skill bridge copies allowlisted staged skill files into a project .claude/skills directory, which Claude can load as agent instructions.
- The bridge can recursively delete the corresponding active skill directory when its exported delete function is called.
- package.json has no preinstall, install, postinstall, or other lifecycle hook.
- The bridge is an exported function with a caller-supplied workspace root; importing it does not perform the write.
- The bridge rejects traversal-like paths and only permits SKILL.md, schema.json, or validated template paths.
- No self-dependency, dynamic code execution, or network activity is involved in the skill bridge.
Source & flagged code
4 flagged · loading sourcePackage metadata claims a different repository identity while copied source loads a runtime dependency bridge.
dist/discovery-2TvulIz8.cjsView on unpkg · L1451Findings
Affected versions and remediation
This report applies to @mulmoclaude/core@4.8.0.
See all versions of @mulmoclaude/core for other recorded verdicts.
- Review the evidence and your use of @mulmoclaude/core@4.8.0 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.