osborn@0.9.186
Voice AI coding assistant - local agent that connects to Osborn frontend
AI Security Review
scanned 1mo ago · by lpm-firewall-aiNo concrete malicious npm lifecycle behavior was found, but the user-triggered remote installer and permission-bypassing agent subprocess create a meaningful dangerous capability. Downgrade to a warning rather than block publication.
Decision evidence
public snapshot- The running agent can download and execute a remote installer when an IDE-start request finds code-server missing.
- It passes the complete process environment to a headless Claude session configured to bypass permissions.
- The agent reads Claude session data from the user home directory and can upload selected research artifacts to a configured frontend.
- The manifest contains only a prepublish hook, so this behavior is not automatically triggered during npm installation.
- No preinstall, install, or postinstall lifecycle hook is present.
- The remote installer is reached from the explicit start_ide runtime path, not package installation.
- The OpenRouter request uses the configured API key as authentication for the declared assistant service.
Source & flagged code
8 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@0.9.186.
See all versions of osborn for other recorded verdicts.
- Review the evidence and your use of osborn@0.9.186 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.