baldart 4.34.0 → 4.34.2

package/CHANGELOG.md

@@ -5,6 +5,23 @@ All notable changes to BALDART will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [4.34.2] - 2026-06-13
+
+**`/new` workflow-delegation gates hardened to BINARY/imperative — close the "ran inline by judgment" escape hatch.** A real `/new` run (mayo, single-card CHORE-0007, BALDART 4.34.1 with both workflows linked + the `Workflow` tool available) ran **both** review-delegation gates inline instead of delegating: the orchestrator invented escape criteria absent from the gate ("single-card → marginal context-economy benefit", "memory note on workflow degeneration") and mis-cited the `feedback_dynamic_workflows_fit` guidance, which actually names the Final Review fan-out as the canonical workflow fit. The degeneration reports it conflated concern the `new2`/`new2-resolve` whole-batch host — a different workflow. Net effect: the run validated only the inline fallback and gave **zero** signal on the `new-card-review` (v4.34.0) path. Both gates were already worded as `IF tool+script → delegate / ELSE inline`, but the prose left enough softness for the model to insert discretion. **PATCH** (gate-prose hardening; no behavior change to the install layout or the workflows themselves — the delegated path was always the intended one).
+
+### Changed
+
+- **`framework/.claude/skills/new/references/review-cycle.md`** — Phase 2.5x branch reframed as a "MECHANICAL, BINARY decision" with an explicit **no-discretion** rule: the soft factors (N=1, small diff, marginal benefit, "a workflow once degenerated", "inline is safer") are named and declared **NOT inputs to the gate**; the `new2` degeneration conflation is corrected inline; running inline while the conditions held is now defined as a **gate violation** with a mandated telemetry log line.
+- **`framework/.claude/skills/new/references/final-review.md`** — Step F.1.5 branch given the same imperative treatment (MUST-delegate, no-discretion rule, `new2`-conflation correction, gate-violation log line).
+
+## [4.34.1] - 2026-06-12
+
+**`reference-integrity` CI gate: recognize the plumbing carve-out.** v4.33.2 introduced a legitimate `subagent_type: general-purpose` dispatch for the mechanical file-revert agent (the REGISTRY.md plumbing carve-out — mechanical git/file ops, never code authoring), but the CI `reference-integrity` gate's R11 anti-fallback rule banned `general-purpose` **unconditionally**, so `main` went red on the v4.33.2 / v4.34.0 tags (the npm publish workflow is independent and succeeded — both versions are live). The gate now allows `subagent_type: general-purpose` **only** on a dispatch line explicitly marked `plumbing carve-out`; every other `general-purpose` dispatch stays banned (R11 intact, intentionally narrow so it can't be abused as a generic fallback). **PATCH** (CI gate fix; no behavior change to `/new`).
+
+### Changed
+
+- **`scripts/check-reference-integrity.js`** — Check A allows `subagent_type: general-purpose` when the dispatch line carries the literal `plumbing carve-out` marker; otherwise the R11 ban stands (with an error message pointing to the marker + REGISTRY.md).
+
 ## [4.34.0] - 2026-06-12
 
 **`/new` per-card review cluster extracted into the `new-card-review` dynamic workflow — kills the dominant context-cost driver on long epics.** On a long epic (10-12 cards) `/new` accumulated context monotonically: every card ran the full review fan-out (Simplify ×3 + Codex/code-review + qa + the verify/FP pass) *inside the orchestrator*, and each step re-paid the whole growing prefix (cumulative `cache_read`). The driver is **turn-count × prefix**, not subagent output volume. Following the validated `new-final-review` pattern (read-only fan-out hosted in a workflow whose context is discarded), the review cluster now runs OUTSIDE the orchestrator and returns only a compact result.

package/VERSION

@@ -1 +1 @@
-4.34.0
+4.34.2

package/framework/.claude/skills/new/references/final-review.md

@@ -54,11 +54,22 @@ diff and emits findings, it touches no git state and asks the user nothing. That
 makes it a clean fit for a **dynamic workflow**, which runs the fan-out as a
 deterministic script outside this orchestrator's context window.
 
-**Branch (decide once, here):**
+**Branch (decide once, here).** This is a **MECHANICAL, BINARY decision** — check the two
+conditions and act on the result. You have **NO discretion**: if both hold you **MUST** delegate.
+A single-card (N=1) batch, a small diff, "the context-economy benefit is marginal here", "a workflow
+once degenerated on this project", or "inline is safer / I want full control" are **NOT inputs to this
+gate** and **MUST NOT** override it. (Any workflow-degeneration reports concern the `new2`/`new2-resolve`
+whole-batch host — a *different* workflow; the canonical guidance is that this read-only F.2–F.4 fan-out
+is **exactly** where a workflow fits.) Delegating IS the SSOT-sanctioned path — the workflow mirrors this
+prose; the inline branch below exists **only** for installs where the `Workflow` tool is genuinely
+absent, never as a judgment-call alternative. If you run inline while the delegation conditions were met,
+that is a **gate violation**: log it as
+`f.final: GATE VIOLATION — ran inline despite Workflow available + new-final-review.js linked`.
 
 - **IF** the `Workflow` tool is available to you **AND** the script
   `.claude/workflows/new-final-review.js` is present (linked by the framework on
-  Claude-enabled installs) → **delegate F.2–F.4** to it:
+  Claude-enabled installs) → you **MUST delegate F.2–F.4** to it (not optional — see the
+  no-discretion rule above):
 
   ```
   Workflow({ name: 'new-final-review', args: {

package/framework/.claude/skills/new/references/review-cycle.md

@@ -12,13 +12,26 @@ state outside the worktree and asks the user nothing on the happy path. That mak
 biggest source of prefix growth on long epics). E2E (Phase 2.6, human-gated + nests a skill) and the
 doc-review (Phase 3, write-mode, must see the FINAL code) stay in the skill.
 
-**Branch (decide once, here — at the top of the review cluster, AFTER Phase 2.5b AC-Closure):**
+**Branch (decide once, here — at the top of the review cluster, AFTER Phase 2.5b AC-Closure).**
+This is a **MECHANICAL, BINARY decision** — evaluate the three conditions below and act on the result.
+You have **NO discretion**: if the delegation conditions hold you **MUST** delegate. The factors that
+feel like reasons to "just run it inline" — a single-card batch, a small/well-scoped diff, "the
+context-economy benefit is marginal here", "a workflow once degenerated on this project", "inline is
+safer / I have full control" — are **NOT inputs to this gate** and **MUST NOT** override it. (Those
+degeneration reports concern the `new2`/`new2-resolve` whole-batch host, a *different* workflow; the
+canonical guidance is that this read-then-fix fan-out is exactly where a workflow fits.) Delegating IS
+the SSOT-sanctioned path: the workflow mirrors this prose, and the inline fallback exists **only** for
+installs where the `Workflow` tool is genuinely absent — not as a judgment-call alternative. If you ever
+run inline while the delegation conditions were met, that is a **gate violation**: log it explicitly as
+`review-cluster: GATE VIOLATION — ran inline despite Workflow available + new-card-review.js linked + non-trivial`
+so it surfaces in telemetry.
 
 - **IF `IS_TRIVIAL`** (§ "Trivial-card fast-lane", re-confirmed on the committed diff) → do NOT delegate.
   Follow the trivial fast-lane in Phase 2.55 below (inline mechanical gates → Phase 3 doc → commit).
 
 - **ELSE IF** the `Workflow` tool is available **AND** `.claude/workflows/new-card-review.js` is present
-  (linked by the framework on Claude-enabled installs) → **delegate the cluster** to it. First build the
+  (linked by the framework on Claude-enabled installs) → you **MUST delegate the cluster** to it (this
+  branch is not optional — see the no-discretion rule above). First build the
   inputs (reusing the existing deterministic logic — do NOT re-implement it):
   - `qaTier` ← the **Phase 3.5 profile selection** (step 19-21b below: `review_profile` floor →
     `skip`/`light` ⇒ `qaTier:"light"` (qa-sentinel deferred to Final), `deep` **or any Phase 3.7 Step-A

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "baldart",
-  "version": "4.34.0",
+  "version": "4.34.2",
   "description": "Claude Agent Framework - Reusable framework for coordinating AI agents and humans in software projects",
   "bin": {
     "baldart": "./bin/baldart.js"