@oked/claude-code@0.6.25
OKed for Claude Code - zero-code human approval for sensitive actions
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe SDK sends the machine hostname to the OKed backend automatically from the runtime hook path, including before safe actions return. Because this identifying telemetry is not disclosed in the README, classify the package as suspicious telemetry.
Decision evidence
public snapshot- The hook calls heartbeat whenever an API key is configured, before safe and warning actions return locally.
- The heartbeat sends the machine hostname to the configured backend’s `/api/v1/heartbeat` endpoint, defaulting to `api.oked.ai`; it is throttled to once per day.
- The README documents `oked init` installing the project hook and pairing the device, but does not disclose hostname telemetry.
- The package has no install lifecycle hook; its only lifecycle script is `prepublishOnly`.
- The hostname ping goes to the package’s own OKed backend and is described in code as a retention presence signal.
- Hook installation and device pairing occur when the user runs `oked init`.
- Approval requests send tool-call details to the configured approval backend, consistent with this package’s stated purpose.
Source & flagged code
6 flaggedRuntime or CLI source writes behavior-bearing configuration into a user or project AI-agent control surface.
dist/cli.jsView on unpkgFindings
Affected versions and remediation
This report applies to @oked/claude-code@0.6.25.
See all versions of @oked/claude-code for other recorded verdicts.
- Review the evidence and your use of @oked/claude-code@0.6.25 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.