@mthanhlm/autodev 0.3.3 → 0.3.4

package/README.md

@@ -60,6 +60,7 @@ Typical brownfield route:
   -> new project
   -> explore codebase
   -> plan phase
+  -> review phase plan
   -> execute next task
   -> execute next task
   -> review
@@ -92,6 +93,7 @@ project -> track -> phase -> tasks
 
 - `/autodev` remains the single entrypoint
 - A phase is planned into one phase overview plus multiple task files
+- `/autodev` stops after planning so the user can review the phase before any execution starts
 - The phase keeps one user-facing orchestration session
 - Each task is preferably executed by a fresh worker agent in the background
 - After each task, the worker reports back with files changed, verification, and blockers

package/autodev/bin/autodev-tools.cjs

@@ -507,16 +507,16 @@ function buildRoute(cwd) {
     };
   }
 
-  const nextExecute = phases.find(phase => phase.planExists && !phase.summaryExists);
-  if (nextExecute) {
+  const nextPlannedReview = phases.find(phase => phase.planExists && !phase.summaryExists);
+  if (nextPlannedReview) {
     return {
-      kind: 'execute_phase',
+      kind: 'plan_review',
       command: '/autodev',
-      manualCommand: `/autodev-execute-phase ${nextExecute.number}`,
-      reason: 'phase_planned_but_not_executed',
+      manualCommand: `/autodev-execute-phase ${nextPlannedReview.number}`,
+      reason: 'phase_planned_awaiting_user_review',
       projectType,
       trackSlug: activeTrack,
-      phaseNumber: nextExecute.number
+      phaseNumber: nextPlannedReview.number
     };
   }
 
@@ -702,7 +702,15 @@ function initPayload(cwd, mode, requestedPhase) {
   const activeTrack = readActiveTrack(cwd);
   const tracks = listTracks(cwd);
   const track = trackPaths(cwd, activeTrack);
-  const phaseMode = mode === 'review-phase' ? 'review' : mode === 'verify-work' ? 'verify' : mode === 'execute-phase' ? 'execute' : mode === 'plan-phase' ? 'plan' : null;
+  const phaseMode = mode === 'review-phase'
+    ? 'review'
+    : mode === 'verify-work'
+      ? 'verify'
+      : mode === 'execute-phase' || mode === 'review-plan'
+        ? 'execute'
+        : mode === 'plan-phase'
+          ? 'plan'
+          : null;
   const phase = phaseMode && activeTrack ? resolvePhase(cwd, activeTrack, requestedPhase, phaseMode) : null;
   const tasks = phase ? listTasksForPhaseDetails(phase) : [];
   const nextTask = nextExecutableTask(tasks);

package/autodev/workflows/autodev.md

@@ -24,7 +24,8 @@ node "$HOME/.claude/autodev/bin/autodev-tools.cjs" status
    - `init_project`: read `@~/.claude/autodev/workflows/new-project.md`, then perform that workflow.
    - `explore_codebase`: read `@~/.claude/autodev/workflows/explore-codebase.md`, then perform that workflow.
    - `plan_phase`: read `@~/.claude/autodev/workflows/plan-phase.md`, then perform that workflow.
-   - `execute_phase`: read `@~/.claude/autodev/workflows/execute-phase.md`, then perform that workflow, which should advance one task-sized unit in the current phase session by trying a fresh task worker first and falling back cleanly when delegation is unavailable.
+   - `plan_review`: read `@~/.claude/autodev/workflows/review-plan.md`, then perform that workflow so the user can review the phase plan before any execution starts.
+   - `execute_phase`: read `@~/.claude/autodev/workflows/execute-phase.md`, then perform that workflow, which should advance one task-sized unit in the current phase session by trying a fresh worker agent first and falling back cleanly when agent delegation is unavailable.
    - `review_phase`: read `@~/.claude/autodev/workflows/review-phase.md`, then perform that workflow.
    - `verify_phase`: read `@~/.claude/autodev/workflows/verify-work.md`, then perform that workflow.
 

package/autodev/workflows/execute-phase.md

@@ -9,8 +9,8 @@ Execute an active-track phase through task-sized worker runs, keep the main sess
 - Execute one task at a time by default.
 - Do not use waves.
 - Prefer a fresh background worker for each task so the phase session does not bloat.
-- If the `Task` tool is available and specialized agents are supported in this environment, delegate task implementation.
-- If a delegation attempt fails with an environment limitation like `specialized agents aren't available`, treat delegation as unavailable for this run and continue cleanly.
+- If agent delegation is available in this environment, delegate task implementation.
+- If a delegation attempt fails with an environment limitation like `specialized agents aren't available`, treat agent delegation as unavailable for this run and continue cleanly.
 - The main phase session should remain orchestration-focused whenever delegation works.
 - The main phase session may update `.autodev/` state and aggregate phase artifacts.
 </rules>
@@ -48,7 +48,7 @@ node "$HOME/.claude/autodev/bin/autodev-tools.cjs" init execute-phase "$ARGUMENT
    - verification plan
    - any dependency note
 
-7. If the `Task` tool is available, first try to spawn one fresh background worker using `autodev-task-worker` for the selected task.
+7. If agent delegation is available, first try to spawn one fresh background worker using `autodev-task-worker` for the selected task.
    Tell it:
    - phase path
    - task path
@@ -60,7 +60,7 @@ node "$HOME/.claude/autodev/bin/autodev-tools.cjs" init execute-phase "$ARGUMENT
 
 8. If worker delegation is unavailable, do not stop on the raw platform message.
    Treat messages like `specialized agents aren't available` as a capability limitation, then:
-   - tell the user plainly that background task workers are unavailable in this environment
+   - tell the user plainly that worker-agent delegation is unavailable in this environment
    - continue with the same task in the current session
    - keep the same task boundaries
    - still write the task summary before moving on

package/autodev/workflows/help.md

@@ -5,7 +5,7 @@ Lean Claude Code workflow. No automatic commits. No branches. No worktrees. Git
 ## Main Entry
 
 - `/autodev`
-  The normal command. It routes automatically through project setup, codebase mapping, planning, execution, review, verification, and cleanup.
+  The normal command. It routes automatically through project setup, codebase mapping, planning, plan review, execution, review, verification, and cleanup.
 
 ## Manual Commands
 
@@ -18,7 +18,7 @@ Lean Claude Code workflow. No automatic commits. No branches. No worktrees. Git
 - `/autodev-plan-phase [phase]`
   Creates or revises one phase plan plus task files in `.autodev/tracks/<track>/phases/NN-type-name/`.
 - `/autodev-execute-phase [phase]`
-  Orchestrates one phase task-by-task, preferring a fresh worker per task, and writes `TASK-NN-SUMMARY.md` plus the final `NN-SUMMARY.md`.
+  Orchestrates one phase task-by-task, preferring a fresh worker agent per task, and writes `TASK-NN-SUMMARY.md` plus the final `NN-SUMMARY.md`.
 - `/autodev-review-phase [phase]`
   Uses four review agents when available for code quality, security, integration, and polish, then writes `NN-REVIEW.md`.
 - `/autodev-verify-work [phase]`
@@ -56,6 +56,7 @@ Typical brownfield route:
   -> new project
   -> explore codebase
   -> plan
+  -> review plan
   -> execute
   -> review
   -> verify

package/autodev/workflows/plan-phase.md

@@ -77,19 +77,23 @@ Each task file must include:
 9. Update the active track `STATE.md` so it points to:
    - `Current Phase: N`
    - `Current Phase Type: <type>`
-   - `Current Step: planned`
+   - `Current Step: plan_review`
    - `Current Task: 01` if tasks exist, otherwise `none`
-   - `Current Task Status: ready`
+   - `Current Task Status: pending_review`
    - `Next Command: /autodev`
    - current ISO timestamp
 
 10. Update `.autodev/STATE.md` so it points to:
    - `Active Track: <slug>`
-   - `Current Step: planned`
+   - `Current Step: plan_review`
    - `Current Task: 01` if tasks exist, otherwise `none`
-   - `Current Task Status: ready`
+   - `Current Task Status: pending_review`
    - `Next Command: /autodev`
    - current ISO timestamp
 
-11. End with a short summary and direct the user to `/autodev`. Mention `/autodev-execute-phase N` as the manual shortcut.
+11. End with a short summary and stop for review.
+   Tell the user:
+   - the plan and task files are ready for review
+   - `/autodev` will open the review checkpoint before any execution starts
+   - `/autodev-execute-phase N` is only the optional manual bypass after review
 </process>

package/autodev/workflows/review-plan.md

@@ -0,0 +1,53 @@
+<purpose>
+Pause after phase planning so the user can review the generated plan and task breakdown before any execution starts.
+</purpose>
+
+<rules>
+- Do not start task execution automatically from a freshly created plan.
+- Keep this as a control checkpoint for the user.
+- Do not edit repository code in this step unless the user explicitly chooses to revise the plan.
+- Prefer `/autodev` as the entrypoint even at this checkpoint.
+</rules>
+
+<process>
+1. Run:
+
+```bash
+node "$HOME/.claude/autodev/bin/autodev-tools.cjs" init review-plan "$ARGUMENTS"
+```
+
+2. If no planned phase is found, stop and direct the user to `/autodev`.
+
+3. Read:
+   - `.autodev/STATE.md`
+   - the active track `STATE.md`
+   - `NN-PLAN.md`
+   - all `TASK-*.md`
+
+4. Show the user a concise review summary:
+   - phase id and phase goal
+   - task list with dependencies
+   - likely files or areas to change
+   - shared verification plan
+
+5. Ask the user what to do next with `AskUserQuestion`:
+   - execute the next task now
+   - revise the phase plan
+   - stop here
+
+6. If the user chooses to execute now:
+   - read `@~/.claude/autodev/workflows/execute-phase.md`
+   - perform that workflow in the same turn
+
+7. If the user chooses to revise the plan:
+   - read `@~/.claude/autodev/workflows/plan-phase.md`
+   - revise the same phase in the same turn
+
+8. If the user chooses to stop here:
+   - keep both state files on:
+     - `Current Step: plan_review`
+     - `Next Command: /autodev`
+   - end with a short note that execution has not started yet
+
+9. End with a short outcome summary and `/autodev` as the default next command.
+</process>

package/commands/autodev/execute-phase.md

@@ -11,7 +11,7 @@ allowed-tools:
   - Glob
   - TodoWrite
   - AskUserQuestion
-  - Task
+  - Agent
 ---
 <objective>
 Execute one active-track phase through task-sized worker runs and record both task-level and phase-level summaries.

package/commands/autodev/explore-codebase.md

@@ -9,7 +9,7 @@ allowed-tools:
   - Grep
   - Glob
   - AskUserQuestion
-  - Task
+  - Agent
 ---
 <objective>
 Explore the current repository, use parallel codebase agents when available, and write the brownfield map into `.autodev/codebase/`.

package/commands/autodev/index.md

@@ -11,7 +11,7 @@ allowed-tools:
   - Glob
   - TodoWrite
   - AskUserQuestion
-  - Task
+  - Agent
   - WebFetch
 ---
 <objective>

package/commands/autodev/review-phase.md

@@ -9,7 +9,7 @@ allowed-tools:
   - Bash
   - Grep
   - Glob
-  - Task
+  - Agent
 ---
 <objective>
 Run the automatic review bundle for one executed phase and write a single review artifact.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mthanhlm/autodev",
-  "version": "0.3.3",
+  "version": "0.3.4",
   "description": "A lean Claude Code workflow system with a single entrypoint, task-based phase execution, and read-only git.",
   "bin": {
     "autodev": "bin/install.js"