@agimon-ai/po-container-runtime-linux-x64@0.0.1-alpha.23
Read-only Po server payload for Linux x64 sandbox containers, mounted at /opt/po and run with Node 24.
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe manifest has no install hook, and the inspected runtime behavior aligns with a container server, OpenAI sign-in, and configurable telemetry. The flagged bundle paths showed local service checks, runtime configuration, and bundled tooling; no concrete attack behavior was found.
Decision evidence
public snapshot- package.json contains only a build script; the container manifest selects the packaged server entrypoint.
- The server wrapper registers session-provided tools and applies their declared restrictions within the session lifecycle.
- The sign-in flow sends OAuth requests to auth.openai.com; it does not show an unrelated credential recipient.
- Telemetry can be disabled through environment settings and stays inactive when no endpoint is resolved and file fallback is off.
- The inspected runtime chunks show local service health checks, configured home and project paths, bundled compiler code, model identifiers, terminal input, and background sync behavior.
Source & flagged code
9 flaggedThe scanner resource watchdog stopped semantic analysis before full coverage; route this package to AI without publishing a static verdict.
package.json contains only a build script; the container manifest selects the packaged server entrypoint.
package.jsonView on unpkg · L37Findings
Affected versions and remediation
This report applies to @agimon-ai/po-container-runtime-linux-x64@0.0.1-alpha.23.
See all versions of @agimon-ai/po-container-runtime-linux-x64 for other recorded verdicts.
- Review the evidence and your use of @agimon-ai/po-container-runtime-linux-x64@0.0.1-alpha.23 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.