@cotal-ai/connector-jcode@0.132.5
Cotal connector for Jcode: a host-mode mesh peer driven through Jcode's Harness API bridge.
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe flagged process and credential flows are consistent with the package’s stated purpose: starting a private Jcode instance to drive an agent session. The inspected source shows selected credentials being made available to Jcode, but no evidence that this package sends them to an unrelated host or performs unauthorized actions.
Decision evidence
public snapshot- The package entry exports the connector extension; its declared scripts are build, typecheck, and test.
- The extension describes its purpose as launching a private Jcode instance and driving one session through Jcode’s Harness API.
- The Jcode SDK shares only selected external CLI credential files into the private JCODE_HOME because Jcode reads them directly.
- The host launches a local Jcode api-bridge and requires a managed control endpoint before running.
- Credential inheritance is enabled by default and includes credentials from other AI CLI tools, so the connector has access to those credentials while running.
- The source also contains process execution and network code, but the inspected execution path ties process launch to the connector’s local Jcode bridge; no unrelated credential transmission or attack behavior was identified.
Source & flagged code
6 flaggedA single source file combines environment access, network access, and code or shell execution; review context before blocking.
dist/host.jsView on unpkg · L49Findings
Affected versions and remediation
This report applies to @cotal-ai/connector-jcode@0.132.5.
See all versions of @cotal-ai/connector-jcode for other recorded verdicts.
- Review the evidence and your use of @cotal-ai/connector-jcode@0.132.5 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.