[verification] MEDIUM: npx @claude-flow/cli@alpha doctor full health check times out

[ruvnet]

Summary

npx @claude-flow/cli@alpha doctor (full health check with no component filter) hangs for >60s and is killed by SIGTERM. The targeted doctor --component federation succeeds in ~64s (ONNX downloads included). The hang blocks operators from using doctor to diagnose system health.

Impact

Operators cannot run full system diagnostics via the CLI. Any automated health-check pipeline using doctor without a component filter will time out and report failure.

Expected

doctor completes with no fail status within a reasonable time (<30s)

Actual

$ timeout 60 npx -y @claude-flow/cli@alpha doctor
[60s pass]
Terminated (exit 143)

The component-scoped run works:

$ timeout 90 npx -y @claude-flow/cli@alpha doctor --component federation
Loading ONNX model: all-MiniLM-L6-v2...
...
✓ Federation Breaker: Federation plugin not installed (optional) — install only if you need cross-installation peering
Summary: 1 passed — All checks passed! System is healthy.

Likely Root Cause

Same root cause as the --version hang (see related HIGH issue): ONNX model download blocks early in the CLI startup path before the full doctor command can begin parallel health checks. With the component filter, the check list is shorter and returns before a global timeout.

Recommended Fix

1. Lazy-load ONNX embeddings — only initialize when a command actually requires them.
2. Add a --timeout <ms> flag to doctor so operators can control the deadline.
3. Add a --no-embeddings flag to skip ML init for lightweight environments.

Related

• HIGH issue: @claude-flow/cli@alpha --version hangs and times out (filed same session)

Environment

• Commit: 28c81c03e3e84555a9238b3217b9f586fc0c7dbc
• Node: v22.22.2
• npm: 10.9.7
• OS: Linux x86_64
• Timestamp UTC: 2026-06-15T19:04:03Z