@mutmutco/claude-plugin@4.5.101
MMI workflow skills and org gates delivery.
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe runtime authentication path sends a saved GitHub credential to the package-selected Hub service without recipient-specific consent. This is active credential disclosure, so the package should be blocked even though the root manifest has no install hook.
Decision evidence
public snapshot- The required agent-configuration location is part of a path-filter helper; the nearby code classifies paths rather than writing agent configuration.
- The root manifest declares no npm lifecycle scripts, so the credential transfer requires runtime CLI activity.
- The CLI reads GitHub credentials from GH_TOKEN or GITHUB_TOKEN and can obtain a saved token through gh auth token.
- The CLI uses a package-selected Hub URL by default, with MMI_HUB_URL as an override.
- Without an MMI Hub token or cached session, the Hub session exchange sends the GitHub token as a bearer credential to /auth/session.
- Hub API header creation invokes the session exchange without a recipient-specific consent check in that path.
Source & flagged code
7 flagged · loading sourceSource appears to send environment or credential material to an external endpoint.
cli/dist/main.cjsView on unpkg · L39This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
cli/dist/main.cjsView on unpkgFindings
Affected versions and remediation
This report applies to @mutmutco/claude-plugin@4.5.101.
See all versions of @mutmutco/claude-plugin for other recorded verdicts.
- Avoid installing @mutmutco/claude-plugin@4.5.101. 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.