@mnemom/mnemom@0.17.28
Transparent AI agent tracing
AI Security Review
scanned 3h ago · by lpm-firewall-aiThe static flags point to Remote Control proxy and certificate-handling code, but source shows the proxy is launched by an explicit CLI option, listens on loopback, and forwards messages to a configured HTTPS gateway. The manifest has no install hook, and the reviewed dependencies and runtime integration show no separate concrete attack behavior.
Decision evidence
public snapshot- package.json has no install lifecycle hook; prepublishOnly runs the build and the package declares CLI entrypoints.
- The prepared policy-engine, chalk, and commander manifests and entry files show their declared package identities and ordinary exported functionality.
- The Slack channel code formats incoming message context for a Claude channel notification; this matches its integration purpose.
- Remote Control starts through the interactive `--remote-control` CLI path and launches the proxy; this behavior is user-invoked.
- The proxy binds to loopback, limits TLS termination to api.anthropic.com, and requires an HTTPS gateway URL from configuration.
- The proxy routes message requests to that gateway and forwards other hosts upstream; its certificate module checks certificate compatibility.
Source & flagged code
5 flagged · loading sourceA single source file combines environment access, network access, and code or shell execution; review context before blocking.
dist/rc-proxy.mjsView on unpkg · L87Findings
Affected versions and remediation
This report applies to @mnemom/mnemom@0.17.28.
See all versions of @mnemom/mnemom for other recorded verdicts.
- Review the evidence and your use of @mnemom/mnemom@0.17.28 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.