brokre
MCP launcher for brokre — local credential broker for Cursor, Claude Code, Kimi Code, Trae, OpenClaw, Hermes Agent, ChatClaw, and other MCP clients
This report shows the evidence for brokre@0.2.32. 2 of 4 known versions are blocked or flagged. The newest version, 0.2.36, is flagged. See all versions.
AI Security Review
scanned 15d ago · by lpm-firewall-aiThe package performs unconsented install-time writes to global AI-client MCP configuration and registers a command that resolves brokre@latest. This is a concrete AI-agent control-surface hijack under the install-control policy.
Decision evidence
public snapshot- The npm postinstall hook runs postinstall.js automatically.
- The hook invokes setup-mcp.js, which writes configuration for detected MCP clients.
- The setup adds brokre to global MCP configuration without a consent prompt.
- The registered command runs brokre@latest through npx, enabling later retrieval of a changing release.
- The setup detects Cursor and other client configuration locations.
- The setup skips clients it does not detect and preserves existing entries unless forced.
- The package provides environment-variable opt-outs for its binary download and MCP setup.
Source & flagged code
6 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time source downloads a native archive from a fixed external host without transport verification, extracts it, and installs an executable payload.
package.jsonView on unpkg · L4Findings
Affected versions and remediation
This report applies to brokre@0.2.32.
The versions list below shows the verdict for each recorded version.
- Avoid installing brokre@0.2.32. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.