feat(models): add reasoning budgets and upstream prompt pruning

docs/upstream-review-budget-and-pruning.md

@@ -0,0 +1,62 @@
+# Budget, prompt pruning, and verification reuse
+
+This change ports the reasoning-budget selection from
+[pstack #366](https://github.com/cursor/plugins/pull/366), prompt pruning from
+[#414](https://github.com/cursor/plugins/pull/414) and
+[#419](https://github.com/cursor/plugins/pull/419), and the shipping build-output
+exception present at
+[`12d587df`](https://github.com/cursor/plugins/commit/12d587dfb20741cafc376c42c696c5f6e2a64487).
+It builds on the priority workflow ports. Source pins remain unchanged because
+this is a selective adaptation, not a full upstream sync.
+
+## Reasoning budget
+
+`model-budget.mjs` previews or applies effort-suffix rewrites using a JSON array
+of model names detected by the selected Harness. `unlimited` keeps current
+efforts. `large`, `medium`, and `small` target `xhigh`, `high`, and `medium`.
+The command preserves model stem, version, trailing `-fast`, and aliases. When
+the target variant is missing, it selects the highest detected lower effort
+of that exact stem, rather than substituting a different model.
+
+The command writes only the selected Harness's overrides and budget label.
+It preserves defaults, other Harnesses, and unrelated fields. Unresolved
+models or duplicate reviewer entries reject the whole write. Repeating an
+already-applied mapping does not rewrite the configuration. `unlimited` does
+not restore an earlier effort level; choosing a larger bounded budget can
+raise the effort to a detected variant.
+
+A model without an effort suffix needs an explicit model choice or the
+Harness's separate reasoning setting. Aliases are preserved, so their effort
+remains under the Harness's control. The budget label is not a spending cap.
+No provider defaults or automatic model-family fallback are introduced.
+
+## Prompt pruning
+
+The port removes repeated review, verification, sequencing, and writing
+instructions identified by upstream as unnecessary for Opus 5.5. The changes
+cover blast-radius, figure-it-out, interrogate references, feature and bug-fix
+playbooks, refactoring, pause-safely, TDD, writing skills, and the affected
+principles. Reflect's fixed finding count was removed in the priority port.
+
+The canonical instructions are shorter for all supported Harnesses, not only
+Opus. Runtime proof requirements, evidence-based findings, human authorization
+boundaries, and source integrity checks remain. No behavioral evaluation of
+other models establishes that the removed reminders are redundant for them;
+that is a review risk of accepting this shared-tree change.
+
+## Shipping exception
+
+Shipping links a procedure for reusing individual lane results after changes
+limited to tests, docs, or lint config. Two builds at the verdict SHA establish
+noise; one at the current head establishes the comparison. Each accepted
+difference needs evidence. A substantive difference in a noisy file still
+requires a rerun. Dev servers, missing saved build output, runtime inputs, and
+inconclusive comparisons do not qualify. CI, mergeability, and review run fresh.
+
+Autopilot verdict language points to that rule rather than requiring an
+unchanged patch ID in every case. Orchestrate's SHA-keyed ledger remains
+strict; it has no build-output reuse implementation.
+
+The exception is an agent-facing procedure, not an automatic classifier.
+Repository tests prove model-budget CLI behavior and packaging integrity, not
+that a model correctly classifies build noise or follows the shortened prompts.

docs/upstream-review-priority.md

@@ -0,0 +1,46 @@
+# Priority workflow ports
+
+This change selectively ports workflow fixes from pstack at
+[`12d587df`](https://github.com/cursor/plugins/commit/12d587dfb20741cafc376c42c696c5f6e2a64487).
+The source pins remain unchanged because model defaults, prompt pruning, and
+verification reuse belong to a separate review. Target digests record these
+reviewed adaptations against the existing pinned sources.
+
+## Owner lifecycle
+
+An autopilot owner's full-lifecycle brief authorizes babysitting. Ordinary
+PR-opening workers still return without starting a watch. Independent
+babysitters report conflicts to the topology owner. Autopilot-full owners
+control their own branches; the Autopilot-stack root controls stack topology.
+
+A code-ready report starts independent verification while self-proof, CI, and
+babysitting continue. Merge-ready or STACK-READY reports include receipts and
+the exact SHA. Each patch-changing push starts a new round. Merge preparation
+requires current-head CI and the verdict checks in the shipping playbook.
+
+Owners record delegated children. Replacement work uses isolated write targets
+unless the old writer's access has been revoked. Audit ticks continue until all
+delegated work is finished, and chat updates report only previously unreported
+changes.
+
+## Evidence and logs
+
+Verification results identify the commit SHAs named in the brief. Measurement
+results also identify sample count, sample definition, and execution order.
+Missing attribution causes one retry, then an explicit gap rather than a pass.
+Workers report every proven defect.
+
+Decision trails remain append-only across handoffs. A `start` row identifies a
+new run and the preceding rows it did not write. Each run audits its own
+stretches and corrects inaccurate records with a superseding row. The portable
+Node logger already creates new files exclusively; the upstream Bash header
+append fix does not replace it.
+
+Reflect reviewers report evidence-backed learnings without a required count.
+
+## Verification boundary
+
+The repository tests validate packaging, installation, skill integrity, and
+source baselines. They do not prove that a future model will follow these
+instructions. Review the owner and babysitter boundary together when changing
+this workflow.

package.json

@@ -18,6 +18,7 @@
     "audit-context": "node scripts/audit-context.mjs",
     "check-package": "node scripts/check-package.mjs",
     "check-models": "node scripts/model-config.mjs",
+    "model-budget": "node scripts/model-budget.mjs",
     "check-upstream": "node scripts/check-upstream.mjs",
     "check-harness-policy": "node scripts/check-harness-policy.mjs",
     "skill-baseline": "node scripts/skill-baseline.mjs",
@@ -28,7 +29,7 @@
     "install-skills": "node scripts/install.mjs",
     "optimize-context": "node scripts/optimize-context.mjs",
     "reconcile-context": "node scripts/reconcile-context.mjs",
-    "test": "node scripts/validate.mjs && node scripts/skill-baseline.mjs --check && node --test scripts/install.test.mjs scripts/install-migration.test.mjs scripts/migration-integration.test.mjs scripts/skill-discovery.test.mjs scripts/harness-targets.test.mjs scripts/remote-discovery.test.mjs scripts/context.test.mjs scripts/model-config.test.mjs scripts/audit-context.test.mjs scripts/worktree-audit.test.mjs scripts/sync-upstream.test.mjs scripts/runtime.test.mjs scripts/environment.test.mjs scripts/skill-integrity.test.mjs scripts/skill-baseline.test.mjs scripts/version-integrity.test.mjs scripts/agent-format.test.mjs scripts/history.test.mjs scripts/harness-policy.test.mjs scripts/harness-project.test.mjs scripts/check-upstream-adaptations.test.mjs"
+    "test": "node scripts/validate.mjs && node scripts/skill-baseline.mjs --check && node --test scripts/install.test.mjs scripts/install-migration.test.mjs scripts/migration-integration.test.mjs scripts/skill-discovery.test.mjs scripts/harness-targets.test.mjs scripts/remote-discovery.test.mjs scripts/context.test.mjs scripts/model-config.test.mjs scripts/model-budget.test.mjs scripts/audit-context.test.mjs scripts/worktree-audit.test.mjs scripts/sync-upstream.test.mjs scripts/runtime.test.mjs scripts/environment.test.mjs scripts/skill-integrity.test.mjs scripts/skill-baseline.test.mjs scripts/version-integrity.test.mjs scripts/agent-format.test.mjs scripts/history.test.mjs scripts/harness-policy.test.mjs scripts/harness-project.test.mjs scripts/check-upstream-adaptations.test.mjs"
   },
   "bin": {
     "mstack": "scripts/install.mjs",