Cursor Delegation

Cursor Delegation

This skill is the policy and the loop for delegating work to Cursor without a /cursor:* command. Presenting output stays governed by the cursor-result-handling skill; prompt shape and model notes stay governed by the cursor-prompting skill. Do not duplicate either — defer to them.

When to delegate — and when not

• Delegate when parallel speed adds real value: an implementation or scaffolding job that can run while the session continues on something else, or an everyday review of a substantive diff.
• Evidence note (portfolio retro, 2026-08-07): across the first 8 ledger dispositions Cursor was never exercised as a delegate, so its routing guidance rests on less evidence than Codex's — when an increment genuinely fits the profile above, prefer delegating it here over defaulting to local work, so the ledger accumulates the comparison the routing policy needs.
• Keep trivial work local. A rename, a one-file edit, a question Claude can answer directly — spawning a background agent there is overhead, not help.
• Non-transferable context, 2026-08-08 (docs/bench-pilot-2026-08-08.md): a bench pilot measured the CODEX arm matching solo Claude's fix rate on one well-localized single-file bugfix cell while costing ~30% more wall time. No Cursor arm has been benched, so that result transfers no verdict here — the evidence note above stands, and bugfix-shaped increments are among the shapes a future Cursor arm should measure.
• Never delegate when the user explicitly asked Claude to do the work personally.
• Run one delegated job of a class at a time (one task, one review). Before firing a new one, run status --json and confirm nothing of the same class is still active.

Disclosure

• Announce every delegation in one short line when starting it — what is being delegated and that it is going to Cursor. Never silently spawn CLI work.

The loop: fire, await, collect

Fire — start the job in the background:

node "${CLAUDE_PLUGIN_ROOT}/scripts/cursor-companion.mjs" task --background [--write] [--profile deep|fast] [--model <model>] "<prompt>"

• The output line started in the background as <jobId> carries the job id. Record it.
• Add --write only when Cursor is supposed to edit files; it removes the per-command confirmation, so add it deliberately.
• This fire → await → collect loop is for task only. Reviews detach differently — see below.

Await — issue the wait as a background Bash task so the harness notifies on completion instead of blocking the session:

Bash({
  command: `node "${CLAUDE_PLUGIN_ROOT}/scripts/cursor-companion.mjs" status <jobId> --wait --timeout-ms 1800000 --json`,
  description: "Await Cursor job",
  run_in_background: true
})

• On wake, parse the JSON. If waitTimedOut is true and the job is still active, re-issue the same wait instead of assuming failure — a timeout is not a verdict.

Collect — fetch the stored output once the job is terminal:

node "${CLAUDE_PLUGIN_ROOT}/scripts/cursor-companion.mjs" result <jobId> --json

• Present it per the cursor-result-handling skill — including its rule that review findings are never auto-applied: present them, stop, and ask which ones to fix.

Reviews: one background step

review and adversarial-review have no companion-side enqueue: the companion parses --background but always runs the review in the foreground, and no started in the background as <jobId> line is printed. Delegate a review by running the whole command as a background Bash task — the wake IS the collect step, carrying the rendered review on stdout:

Bash({
  command: `node "${CLAUDE_PLUGIN_ROOT}/scripts/cursor-companion.mjs" review [--base <ref>]`,
  description: "Cursor review in background",
  run_in_background: true
})

• adversarial-review [--base <ref>] [focus text] detaches the same way.
• The run still records a tracked job, so status and result work on it afterwards.

Routing

• Fast implementation, scaffolding, everyday review → task or review.
• Leave the model unset by default — Cursor routes auto server-side. There is no --effort flag; depth is a model choice, and --profile deep|fast is a named shortcut for it on task only (review/adversarial-review reject --profile).

Failure honesty

• If the companion errors, or Cursor was never actually invoked, report that plainly and stop. Never substitute Claude-authored output as the delegate's.