Split the six cognitive-complexity hotspots into options → operation → renderer

[Vonng]

Why

Measured cognitive complexity: runNetwork 157, Config.Validate 112, runImage 110, runSetupCommand 100, Manager.Up 97, runPrivateCommand 92. The CLI handlers mix argument checks, host probing, and three output formats in one body, which is why every review finds another bypass of the error layer.

Proposed shape

• Per CLI handler: typed options (mostly present) → one domain call returning a typed result → one renderer per format.
• runNetwork: separate install / uninstall / status handlers, per OS.
• Manager.Up: extract the existing-state branch (diff, phase classification, the startExisting decision) into a named function shared with Reload.
• Config.Validate remains one validator; factor the repeated node loops but do not split it to chase a number.

Acceptance

• No function above roughly 40; behaviour and exit codes unchanged with the existing tests as the oracle; one extraction per commit.
• Best done after Return typed results and errors from the command layer; render and classify only at the root #6 so each renderer has a typed result to consume.

[Vonng]

Post-#6 assessment: defer; issue remains open

The prerequisite is complete: #6 removed runner-owned failure rendering, the stderr reason back-channel, and scattered exit-code policy. Every command family now returns a typed commandOutcome, error, so the correctness risk that originally made these hotspots urgent is materially reduced.

The remaining size is real, however: runNetwork, runPrivateCommand, runImage, runSetupCommand, and Manager.Up still contain substantial dispatch/domain sequencing, while File.Validate remains one intentionally cohesive validator. Meeting the current acceptance of roughly 40 lines for every function would require a new multi-commit extraction wave, not a necessary tail fix, and would encourage artificial splitting of File.Validate contrary to this issue and the cleanup objective.

Decision for the current cleanup: do not select #12 for implementation. Keep it open as a focused follow-up after 0.1.1 behavior settles. A future Plan should split it by command/domain owner, preserve the typed root boundary from #6, and measure whether each extraction removes duplicated policy rather than pursuing a line-count number. No code or close action is attached to this assessment.

[Vonng]

Final status summary — OPEN / deliberately deferred

Why this exists: several orchestration functions remain cognitively large because they combine dispatch, platform/domain sequencing, and result construction. Before #6, that size also allowed output and exit-policy bypasses.

What changed underneath it: #6 completed the typed outcome/error boundary and centralized all rendering and exit decisions. That removes the highest-risk correctness problem, but runNetwork, runPrivateCommand, runImage, runSetupCommand, and Manager.Up remain substantial. File.Validate is intentionally cohesive and should not be split merely to hit a number.

Why it remains open: satisfying the current roughly-40-line acceptance requires a separate extraction program, not a tail cleanup. Future work should proceed one command/domain owner per commit, preserve #6's typed boundary, and demonstrate removal of duplicated policy. No code was committed or issue closed during this cleanup wave.