fix: scope workflow commits to defined commit points

.claude-plugin/marketplace.json

@@ -12,7 +12,7 @@
       "name": "dev-workflows",
       "source": "./dev-workflows",
       "strict": true,
-      "version": "0.25.5",
+      "version": "0.25.6",
       "description": "Skills + Subagents for backend development - Use skills for coding guidance, or run recipe workflows for full orchestrated agentic coding with specialized agents",
       "author": {
         "name": "Shinsuke Kagawa",
@@ -83,7 +83,7 @@
       "name": "dev-workflows-frontend",
       "source": "./dev-workflows-frontend",
       "strict": true,
-      "version": "0.25.5",
+      "version": "0.25.6",
       "description": "Skills + Subagents for React/TypeScript - Use skills for coding guidance, or run recipe workflows for full orchestrated agentic coding with specialized agents",
       "author": {
         "name": "Shinsuke Kagawa",
@@ -157,7 +157,7 @@
       "name": "dev-workflows-fullstack",
       "source": "./dev-workflows-fullstack",
       "strict": true,
-      "version": "0.25.5",
+      "version": "0.25.6",
       "description": "Skills + Subagents for fullstack development (backend + React/TypeScript) - Use skills for coding guidance, or run recipe workflows for full orchestrated agentic coding with specialized agents",
       "author": {
         "name": "Shinsuke Kagawa",
@@ -247,7 +247,7 @@
       "name": "dev-skills",
       "source": "./dev-skills",
       "strict": true,
-      "version": "0.25.5",
+      "version": "0.25.6",
       "description": "Lightweight skills for users with existing workflows - coding best practices, testing principles, and design guidelines without recipe workflows or agents",
       "author": {
         "name": "Shinsuke Kagawa",

dev-skills/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "dev-skills",
   "description": "Lightweight skills for users with existing workflows - coding best practices, testing principles, and design guidelines without recipe workflows or agents",
-  "version": "0.25.5",
+  "version": "0.25.6",
   "author": {
     "name": "Shinsuke Kagawa",
     "url": "https://github.com/shinpr"

dev-workflows-frontend/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "dev-workflows-frontend",
   "description": "Skills + Subagents for React/TypeScript - Use skills for coding guidance, or run recipe workflows for full orchestrated agentic coding with specialized agents",
-  "version": "0.25.5",
+  "version": "0.25.6",
   "author": {
     "name": "Shinsuke Kagawa",
     "url": "https://github.com/shinpr"

dev-workflows-frontend/skills/recipe-front-build/SKILL.md

@@ -20,11 +20,11 @@ Before the first finding disposition, read `references/review-resolution.md` fro
 
 **Execution Protocol**:
 1. **Invoke named specialists for deliverable production** — pass deliverable paths between them and validate their results (see subagents-orchestration-guide "Orchestrator Execution Boundary")
-2. **Follow the 4-step task cycle exactly**: execute → branch on executor result → quality-fix → commit
+2. **Follow the 4-step task cycle exactly for each task in the Consumed Task Set**: execute → branch on executor result → quality-fix → commit. Corrections produced outside that cycle reuse its executor-result branching and quality gate, and defer only the commit to their own phase's rule
 3. **Enter autonomous mode** when the user provides execution instruction with an existing Work Plan or task files — this IS the batch approval
 4. **Scope**: Complete consumed task-set execution, post-implementation verification, consumed-task cleanup, and completion reporting in order, or stop autonomous execution at the current phase for a valid user-owned escalation. Advance only when the current phase's stated transition condition is satisfied.
 
-**CRITICAL**: Run quality-fixer-frontend before every commit.
+**CRITICAL**: Commit only after quality-fixer-frontend returns `approved` or `verification_incomplete`. A quality-fixer-frontend pass authorizes a commit at a defined commit point; it does not create one.
 
 Work plan: $ARGUMENTS
 
@@ -81,7 +81,7 @@ Recompute the Consumed Task Set using the same restricted pattern from the Consu
   - Other environments (tests, quality tools) → Quality agents retain proof limitations while the task cycle continues
 
 ## Task Execution Cycle (4-Step Cycle)
-**MANDATORY EXECUTION CYCLE**: `execute → branch on executor result → quality-fix → commit`
+**MANDATORY EXECUTION CYCLE** (per task in the Consumed Task Set): `execute → branch on executor result → quality-fix → commit`
 
 For EACH task in the Consumed Task Set, YOU MUST:
 1. **EXECUTE**: invoke Agent tool (subagent_type: "dev-workflows-frontend:task-executor-frontend") → Record the current HEAD as `diffBase`, pass `task_file: [path]`, and receive the structured response
@@ -117,7 +117,7 @@ Apply subagents-orchestration-guide's Post-Implementation Review status-routing
 
 ## Final Cleanup
 
-Before the completion report, delete the implementation task files this recipe consumed. Their work is committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
+Before the completion report, commit the post-review corrections applied at Review Resolution convergence when any remain uncommitted, applying subagents-orchestration-guide Commit Boundary Check, then delete the implementation task files this recipe consumed. Their work is then committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
 
 - Delete every file in the Consumed Task Set
 - Preserve the work plan itself (`docs/plans/{plan-name}.md`) — the user decides whether to delete it after final review

dev-workflows-frontend/skills/subagents-orchestration-guide/SKILL.md

@@ -195,7 +195,7 @@ For Small, execute one direct-scope 4-step cycle. Complete after `approved`, or
 
 Reviewer findings are candidates. Create correction work only from the Review Resolution `apply` set.
 
-**Fix-cycle handoff**: Apply Review Resolution and invoke each correction owner it selects. For an author-owned technical-artifact correction, invoke the layer-appropriate technical designer in update mode, run the artifact's existing document-reviewer and applicable design-sync gates, then re-run the originating reviewer. For an executor-owned correction, invoke the layer-appropriate executor with its original `task_file` or direct-scope fields plus `correction_findings` as the complete `apply` finding objects verbatim with only their dispositions added, then run the applicable quality gate. When both owners are required, Review Resolution's author-first re-evaluation controls the order. Carry `prior_feedback` only to reconciliation reviewers.
+**Fix-cycle handoff**: Apply Review Resolution and invoke each correction owner it selects. For an author-owned technical-artifact correction, invoke the layer-appropriate technical designer in update mode, run the artifact's existing document-reviewer and applicable design-sync gates, then re-run the originating reviewer. For an executor-owned correction, invoke the layer-appropriate executor with its original `task_file` or direct-scope fields plus `correction_findings` as the complete `apply` finding objects verbatim with only their dispositions added, then branch on the executor result through the per-task cycle's step 2, including its conditional integration-test-reviewer path, and run the applicable quality gate. When both owners are required, Review Resolution's author-first re-evaluation controls the order. Carry `prior_feedback` only to reconciliation reviewers. Post-implementation corrections stay uncommitted through this cycle: the reviewers read the current working tree, so run the applicable quality gate and re-run the originating reviewer on the uncommitted changes, and commit the applied corrections once through Commit Boundary Check after the complete review set reaches Review Resolution convergence.
 
 **Re-run rule**: After any applied post-implementation correction, re-run each reviewer with at least one correction applied from its latest result. Retain any other reviewer result completed by Post-Implementation Review Status Routing or Review Resolution only when repository evidence establishes that the correction preserved its review boundary; otherwise re-run that reviewer. After Specialist Result Acceptance recovers a blocked review prerequisite, re-run that reviewer. Review Resolution convergence governs acceptance and preserves resolved declines.
 

dev-workflows-fullstack/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "dev-workflows-fullstack",
   "description": "Skills + Subagents for fullstack development (backend + React/TypeScript) - Use skills for coding guidance, or run recipe workflows for full orchestrated agentic coding with specialized agents",
-  "version": "0.25.5",
+  "version": "0.25.6",
   "author": {
     "name": "Shinsuke Kagawa",
     "url": "https://github.com/shinpr"

dev-workflows-fullstack/skills/recipe-build/SKILL.md

@@ -20,11 +20,11 @@ Before the first finding disposition, read `references/review-resolution.md` fro
 
 **Execution Protocol**:
 1. **Invoke named specialists for deliverable production** — pass deliverable paths between them and validate their results (see subagents-orchestration-guide "Orchestrator Execution Boundary")
-2. **Follow the 4-step task cycle exactly**: execute → branch on executor result → quality-fix → commit
+2. **Follow the 4-step task cycle exactly for each task in the Consumed Task Set**: execute → branch on executor result → quality-fix → commit. Corrections produced outside that cycle reuse its executor-result branching and quality gate, and defer only the commit to their own phase's rule
 3. **Enter autonomous mode** when the user provides execution instruction with an existing Work Plan or task files — this IS the batch approval
 4. **Scope**: Complete consumed task-set execution, post-implementation verification, consumed-task cleanup, and completion reporting in order, or stop autonomous execution at the current phase for a valid user-owned escalation. Advance only when the current phase's stated transition condition is satisfied.
 
-**CRITICAL**: Run quality-fixer before every commit.
+**CRITICAL**: Commit only after quality-fixer returns `approved` or `verification_incomplete`. A quality-fixer pass authorizes a commit at a defined commit point; it does not create one.
 
 Work plan: $ARGUMENTS
 
@@ -81,7 +81,7 @@ Recompute the Consumed Task Set using the same restricted pattern from the Consu
   - Other environments (tests, quality tools) → Quality agents retain proof limitations while the task cycle continues
 
 ## Task Execution Cycle (4-Step Cycle)
-**MANDATORY EXECUTION CYCLE**: `execute → branch on executor result → quality-fix → commit`
+**MANDATORY EXECUTION CYCLE** (per task in the Consumed Task Set): `execute → branch on executor result → quality-fix → commit`
 
 For EACH task in the Consumed Task Set, YOU MUST:
 1. **EXECUTE**: invoke Agent tool (subagent_type: "dev-workflows-fullstack:task-executor") → Record the current HEAD as `diffBase`, pass `task_file: [path]`, and receive the structured response
@@ -117,7 +117,7 @@ Apply subagents-orchestration-guide's Post-Implementation Review status-routing
 
 ## Final Cleanup
 
-Before the completion report, delete the implementation task files this recipe consumed. Their work is committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
+Before the completion report, commit the post-review corrections applied at Review Resolution convergence when any remain uncommitted, applying subagents-orchestration-guide Commit Boundary Check, then delete the implementation task files this recipe consumed. Their work is then committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
 
 - Delete every file in the Consumed Task Set
 - Preserve the work plan itself (`docs/plans/{plan-name}.md`) — the user decides whether to delete it after final review

dev-workflows-fullstack/skills/recipe-front-build/SKILL.md

@@ -20,11 +20,11 @@ Before the first finding disposition, read `references/review-resolution.md` fro
 
 **Execution Protocol**:
 1. **Invoke named specialists for deliverable production** — pass deliverable paths between them and validate their results (see subagents-orchestration-guide "Orchestrator Execution Boundary")
-2. **Follow the 4-step task cycle exactly**: execute → branch on executor result → quality-fix → commit
+2. **Follow the 4-step task cycle exactly for each task in the Consumed Task Set**: execute → branch on executor result → quality-fix → commit. Corrections produced outside that cycle reuse its executor-result branching and quality gate, and defer only the commit to their own phase's rule
 3. **Enter autonomous mode** when the user provides execution instruction with an existing Work Plan or task files — this IS the batch approval
 4. **Scope**: Complete consumed task-set execution, post-implementation verification, consumed-task cleanup, and completion reporting in order, or stop autonomous execution at the current phase for a valid user-owned escalation. Advance only when the current phase's stated transition condition is satisfied.
 
-**CRITICAL**: Run quality-fixer-frontend before every commit.
+**CRITICAL**: Commit only after quality-fixer-frontend returns `approved` or `verification_incomplete`. A quality-fixer-frontend pass authorizes a commit at a defined commit point; it does not create one.
 
 Work plan: $ARGUMENTS
 
@@ -81,7 +81,7 @@ Recompute the Consumed Task Set using the same restricted pattern from the Consu
   - Other environments (tests, quality tools) → Quality agents retain proof limitations while the task cycle continues
 
 ## Task Execution Cycle (4-Step Cycle)
-**MANDATORY EXECUTION CYCLE**: `execute → branch on executor result → quality-fix → commit`
+**MANDATORY EXECUTION CYCLE** (per task in the Consumed Task Set): `execute → branch on executor result → quality-fix → commit`
 
 For EACH task in the Consumed Task Set, YOU MUST:
 1. **EXECUTE**: invoke Agent tool (subagent_type: "dev-workflows-fullstack:task-executor-frontend") → Record the current HEAD as `diffBase`, pass `task_file: [path]`, and receive the structured response
@@ -117,7 +117,7 @@ Apply subagents-orchestration-guide's Post-Implementation Review status-routing
 
 ## Final Cleanup
 
-Before the completion report, delete the implementation task files this recipe consumed. Their work is committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
+Before the completion report, commit the post-review corrections applied at Review Resolution convergence when any remain uncommitted, applying subagents-orchestration-guide Commit Boundary Check, then delete the implementation task files this recipe consumed. Their work is then committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
 
 - Delete every file in the Consumed Task Set
 - Preserve the work plan itself (`docs/plans/{plan-name}.md`) — the user decides whether to delete it after final review

dev-workflows-fullstack/skills/recipe-fullstack-implement/SKILL.md

@@ -100,7 +100,7 @@ When user responds to questions:
   - Other environments (tests, quality tools) → Quality agents retain proof limitations while the task cycle continues
 
 **Required Flow Compliance**:
-- Run quality-fixer (layer-appropriate) before every commit
+- Commit only after the layer-appropriate quality-fixer returns `approved` or `verification_incomplete`; a quality-fixer pass authorizes a commit at a defined commit point rather than creating one
 - Obtain user approval before Edit/Write/MultiEdit outside autonomous mode
 
 ## Mandatory Orchestrator Responsibilities
@@ -141,7 +141,7 @@ Apply subagents-orchestration-guide's Post-Implementation Review status-routing
 
 ### Final Cleanup
 
-Before the completion report, delete the implementation task files this recipe consumed. Their work is committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
+Before the completion report, commit the post-review corrections applied at Review Resolution convergence when any remain uncommitted, applying subagents-orchestration-guide Commit Boundary Check, then delete the implementation task files this recipe consumed. Their work is then committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
 
 - Delete every file matching `docs/plans/tasks/{plan-name}-backend-task-*.md` and `docs/plans/tasks/{plan-name}-frontend-task-*.md` (the `{plan-name}` derived from the work plan path used in this run)
 - Preserve the work plan itself (`docs/plans/{plan-name}.md`) — the user decides whether to delete it after final review

dev-workflows-fullstack/skills/recipe-implement/SKILL.md

@@ -91,7 +91,7 @@ After Structural Scale is determined, follow only that scale's applicable path.
   - Other environments (tests, quality tools) → Quality agents retain proof limitations while the task cycle continues
 
 **Required Flow Compliance**:
-- Run quality-fixer before every commit
+- Commit only after quality-fixer returns `approved` or `verification_incomplete`; a quality-fixer pass authorizes a commit at a defined commit point rather than creating one
 - Obtain user approval before Edit/Write/MultiEdit outside autonomous mode
 
 ## Mandatory Orchestrator Responsibilities
@@ -130,7 +130,7 @@ For Small, skip this document-dependent review. Retry a retained verification li
 
 ### Final Cleanup
 
-For Medium/Large, before the completion report, delete the implementation task files this recipe consumed. Their work is committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
+For Medium/Large, before the completion report, commit the post-review corrections applied at Review Resolution convergence when any remain uncommitted, applying subagents-orchestration-guide Commit Boundary Check, then delete the implementation task files this recipe consumed. Their work is then committed; `docs/plans/` is ephemeral working state and is not retained between recipe runs:
 
 - Delete every file matching `docs/plans/tasks/{plan-name}-task-*.md` (the `{plan-name}` derived from the work plan path used in this run)
 - Preserve the work plan itself (`docs/plans/{plan-name}.md`) — the user decides whether to delete it after final review