@mulmoclaude/core@5.11.2
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 14d ago · by lpm-firewall-aiThe registry import path can write behavior-bearing skill files into Claude’s active workspace skills directory, so this is a guarded agent capability. No evidence shows that npm installation triggers it or that the behavior is malicious; classify as suspicious and warn.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
A caller invokes the collection registry import flow.
Impact
Imported skill content becomes available to Claude in that workspace; no malicious instructions or unconsented install-time activation were established.
Mechanism
The flow validates and stages imported skill content, then mirrors allowed files into `.claude/skills/<slug>`.
AI rationale
The registry import path can write behavior-bearing skill files into Claude’s active workspace skills directory, so this is a guarded agent capability. No evidence shows that npm installation triggers it or that the behavior is malicious; classify as suspicious and warn.
Evidence
dist/collection/registry/server/index.jsdist/skill-bridge/index.js.claude/skills/<slug>/
Network endpoints1
asia-northeast1-mulmoserver.cloudfunctions.net
Decision evidence
public snapshotAI called this Suspicious at 83.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- package.json defines build, typecheck, test, and lint scripts, with no install lifecycle hook.
- Registry import validates a bundle, writes it to the selected skill directory, and mirrors allowed files into `.claude/skills/<slug>`.
- The skill bridge copies staged skill files into Claude’s active skills directory; this is a real agent configuration capability.
- Google broker credentials use a package-associated HTTPS endpoint by default; HTTP overrides are restricted to loopback.
Evidence against
- The skill bridge restricts mirrored paths to safe skill files and templates, and validates slugs and staging paths.
- The agent configuration writes occur in registry import or staging flows, not through an npm lifecycle hook.
- No concrete credential theft, destructive behavior, or remote code execution was identified.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetwork
Source & flagged code
4 flagged · loading sourcedist/discovery-CbwHtcwk.jsView file
1447package = @mulmoclaude/core; repositoryIdentity = mulmoclaude; dependency = @duckdb/node-api
L1447: async function duckDbInstance() {
L1448: if (instancePromise === null) instancePromise = import("@duckdb/node-api").then((mod) => mod.DuckDBInstance.create(":memory:"));
L1449: try {
High
Copied Package Dependency Bridge
Package metadata claims a different repository identity while copied source loads a runtime dependency bridge.
dist/discovery-CbwHtcwk.jsView on unpkg · L1447dist/skill-bridge/index.jsView file
•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation.
import { t as issafeactiontemplatepath } from "../templatepath-k_wnbl_q.js";
import { mkdirsync, readfilesync, renamesync, rmsync, writefilesync } from "node:fs";
import path from "node:path";
var data_skills_dir = path.join("data", "skills");
var claude_skills_dir = path.join(".claude", "skills");
var sk[redacted] = "skill.md";
var schema_filename = "schema.json";
var slug_re = /^[a-z0-9]+(-[a-z0-9]+)*$/;
var rm_re = /^\s*rm\s+((?:-[a-za-z]+\s+)+)['"]?data\/skills\/([a-z0-9-]+)\/?['"]?\s*$/;
var recursive_flag_re = /[rr]/;
function dataskilldir(workspaceroot, slug) {
Findings
4 High3 Medium5 Low
HighCopied Package Dependency Bridgedist/discovery-CbwHtcwk.js
HighTrigger Reachable External Ai Agent Control Surface Mutationdist/skill-bridge/index.js
HighSemantic Analysis Limiteddist/promptSafety-yET9jLep.cjs#virtual:normalized:round1
HighPrevious Version Dangerous Deltadist/google/index.cjs
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowTelemetry
LowUrl Strings
Affected versions and remediation
This report applies to @mulmoclaude/core@5.11.2.
See all versions of @mulmoclaude/core for other recorded verdicts.
- Review the evidence and your use of @mulmoclaude/core@5.11.2 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.