osborn@0.9.155
Voice AI coding assistant - local agent that connects to Osborn frontend
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThe code contains a serious runtime data-exposure and remote-installer capability, but source inspection found no automatic install-time execution or covert exfiltration chain. It is suspicious dangerous functionality rather than proven malware.
Decision evidence
public snapshot- The package provides a CLI that launches its runtime with the caller's environment.
- The runtime exposes an archive of Claude projects and skills; authentication is skipped when its optional sync token is unset.
- The API server listens on all interfaces by default, making that export reachable beyond localhost if its port is exposed.
- Session utilities resolve and read Claude data from the user's home configuration directory.
- The runtime can download and execute a remote code-server installer through a shell.
- The manifest has only a prepublish hook, not an install-time lifecycle hook.
- The risky behavior requires an explicit CLI/runtime launch; no automatic registry-install execution was found.
- The reviewed network calls are aligned with the voice, meeting, and IDE features rather than a hidden collector.
Source & flagged code
9 flagged · loading sourcePackage source references dynamic require/import behavior.
dist/jsonl-search.jsView on unpkg · L34Findings
Affected versions and remediation
This report applies to osborn@0.9.155.
See all versions of osborn for other recorded verdicts.
- Review the evidence and your use of osborn@0.9.155 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.