maxsimcli 4.16.0 → 5.0.1

package/dist/assets/templates/workflows/execute-phase.md

@@ -1,752 +0,0 @@
-<sanity_check>
-Before executing any step in this workflow, verify:
-1. The current directory contains a `.planning/` folder — if not, stop and tell the user to run `/maxsim:init` first.
-2. `.planning/ROADMAP.md` exists — if not, stop and tell the user to initialize the project.
-</sanity_check>
-
-<purpose>
-Execute all plans in a phase using wave-based parallel execution. Orchestrator stays lean — delegates plan execution to subagents. Plans are read from GitHub issue comments; summaries are posted as GitHub comments; task sub-issues are transitioned on the board per task.
-</purpose>
-
-<core_principle>
-Orchestrator coordinates, not executes. Each subagent loads the full execute-plan context. Orchestrator: discover plans from GitHub → analyze deps → group waves → spawn agents → handle checkpoints → collect results.
-</core_principle>
-
-<agent_teams>
-**Agent Teams Model:** MAXSIM uses Agent Teams for grouping parallel agents, NOT for peer-to-peer communication.
-
-- `team_name` groups agents in a wave so Claude Code can track them as a unit
-- All coordination routes through the orchestrator (this workflow)
-- Subagents CANNOT spawn other subagents — only the orchestrator spawns agents
-- Inter-wave handoff: orchestrator passes wave N results as context to wave N+1 agents
-- Status tracking: orchestrator maintains a progress table per wave
-</agent_teams>
-
-<required_reading>
-Read STATE.md before any operation to load project context.
-</required_reading>
-
-<process>
-
-<step name="initialize" priority="first">
-Load all context in one call:
-
-```bash
-INIT=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs init execute-phase "${PHASE_ARG}")
-```
-
-Parse JSON for: `executor_model`, `verifier_model`, `commit_docs`, `parallelization`, `branching_strategy`, `branch_name`, `phase_found`, `phase_dir`, `phase_number`, `phase_name`, `phase_slug`, `plans`, `incomplete_plans`, `plan_count`, `incomplete_count`, `state_exists`, `roadmap_exists`, `phase_req_ids`, `phase_issue_number`, `task_mappings`.
-
-Also extract parallel execution fields:
-```
-WORKTREE_MODE (from init JSON worktree_mode — "auto", "always", or "never")
-MAX_PARALLEL_AGENTS (from init JSON max_parallel_agents — default 10)
-REVIEW_CONFIG (from init JSON review_config — spec_review, code_review, simplify_review, retry_limit)
-```
-
-**If `phase_found` is false:** Error — phase directory not found.
-**If `plan_count` is 0:** Check GitHub before reporting no plans (see discover_and_group_plans step).
-**If `state_exists` is false but `.planning/` exists:** Offer reconstruct or continue.
-
-When `parallelization` is false, plans within a wave execute sequentially.
-</step>
-
-<step name="handle_branching">
-Check `branching_strategy` from init:
-
-**"none":** Skip, continue on current branch.
-
-**"phase" or "milestone":** Use pre-computed `branch_name` from init:
-```bash
-git checkout -b "$BRANCH_NAME" 2>/dev/null || git checkout "$BRANCH_NAME"
-```
-
-All subsequent commits go to this branch. User handles merging.
-</step>
-
-<step name="validate_phase">
-From init JSON: `phase_dir`, `plan_count`, `incomplete_count`.
-
-Report: "Found {plan_count} plans in {phase_dir} ({incomplete_count} incomplete)"
-</step>
-
-<step name="discover_and_group_plans">
-
-## Plan Discovery -- GitHub First
-
-When `phase_issue_number` is set (GitHub integration active):
-
-1. **Check for external edits (WIRE-06):**
-   ```
-   mcp_detect_external_edits(issue_number: phase_issue_number)
-   ```
-   If external edits detected: warn user before continuing.
-
-2. **Fetch the phase issue and its comments:**
-   ```
-   mcp_get_issue_detail(issue_number: phase_issue_number)
-   ```
-
-3. **Parse plan comments:** Identify comments containing either:
-   - `<!-- maxsim:type=plan -->` HTML marker, OR
-   - A heading `## Plan NN` (where NN is a number)
-
-4. **For each plan comment:**
-   - Extract plan number from the heading or frontmatter
-   - Parse YAML frontmatter from the comment body for: `wave`, `dependencies`, `autonomous`, `objective`, `gap_closure`, `task_count`
-   - Store plan content in memory as `PLAN_COMMENTS[]`
-
-5. **Determine completion status for each plan:**
-   - Call `mcp_list_sub_issues(issue_number: phase_issue_number)` to get all task sub-issues
-   - A plan is complete when all its task sub-issues are closed (cross-reference using `task_mappings`)
-   - Also check for `<!-- maxsim:type=summary -->` comments as a secondary completion signal
-
-6. **Build plan inventory:**
-   - `plans[]`: each with `id`, `wave`, `autonomous`, `objective`, `has_summary` (true if all sub-issues closed), `plan_comment_body`, `task_mappings`
-   - `waves`: map of wave number → plan IDs
-   - `incomplete`: IDs of plans that are not yet complete
-   - `has_checkpoints`: true if any plan has `autonomous: false`
-
-**Filtering:** Skip plans where `has_summary: true`. If `--gaps-only`: also skip non-gap_closure plans. If all filtered: "No matching incomplete plans" → exit.
-
-**If GitHub integration is NOT active** (`phase_issue_number` is not set):
-
-GitHub Issues is the source of truth for plans. Report the error and exit:
-```
-GitHub integration required. Ensure GitHub Issues is configured: run /maxsim:init to set up.
-```
-
-Report:
-```
-## Execution Plan
-
-**Phase {X}: {Name}** — {total_plans} plans across {wave_count} waves
-
-| Wave | Plans | What it builds |
-|------|-------|----------------|
-| 1 | 01-01, 01-02 | {from plan objectives, 3-8 words} |
-| 2 | 01-03 | ... |
-```
-</step>
-
-<step name="decide_execution_mode">
-After discovering plans and grouping by wave, determine whether to use batch (worktree) or standard execution.
-
-```bash
-# Parse --worktrees / --no-worktrees flags from $ARGUMENTS
-FLAG_OVERRIDE=""
-if echo "$ARGUMENTS" | grep -q "\-\-worktrees"; then FLAG_OVERRIDE="worktrees"; fi
-if echo "$ARGUMENTS" | grep -q "\-\-no-worktrees"; then FLAG_OVERRIDE="no-worktrees"; fi
-
-# Call CLI decision function
-MODE_RESULT=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs decide-execution-mode \
-  "${INCOMPLETE_PLAN_COUNT}" "${WAVE_COUNT}" \
-  --worktree-mode "${WORKTREE_MODE}" \
-  ${FLAG_OVERRIDE:+--flag-override "$FLAG_OVERRIDE"})
-EXECUTION_MODE=$(echo "$MODE_RESULT" | node -e "process.stdin.on('data',d=>console.log(JSON.parse(d).mode))")
-```
-
-Display decision:
-```
-**Execution Mode:** {Batch (worktree isolation) | Standard}
-{If batch: "Will create {N} worktrees for parallel plan execution"}
-{If standard and parallelization true: "Parallel within waves, sequential across waves"}
-{If standard and parallelization false: "Sequential execution"}
-```
-
-**If batch mode, validate plan independence:**
-```bash
-# Build plans JSON from plan inventory data
-VALID=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs validate-plan-independence "$PLANS_JSON")
-```
-If conflicts found: report conflicts, fall back to standard mode with warning.
-</step>
-
-<step name="execute_waves">
-Execute each wave in sequence. The execution path depends on `EXECUTION_MODE` from the decide_execution_mode step.
-
-**BATCH PATH (when EXECUTION_MODE == "batch"):**
-
-For each wave:
-
-1. **Create worktrees for all plans in the wave:**
-   ```bash
-   for PLAN_ID in $WAVE_PLANS; do
-     WT=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs worktree-create "${PHASE_NUMBER}" "$PLAN_ID" "$WAVE_NUM")
-     # Parse WT JSON for path and branch
-   done
-   ```
-
-2. **Describe what is being built (BEFORE spawning):**
-
-   Read each plan's `<objective>` from the plan comment content. Extract what is being built and why.
-
-   ```
-   ---
-   ## Wave {N} (Batch Mode)
-
-   **{Plan ID}: {Plan Name}**
-   {2-3 sentences: what this builds, technical approach, why it matters}
-
-   Spawning {count} isolated agent(s) in worktrees...
-   ---
-   ```
-
-3. **Spawn executor agents with worktree isolation:**
-
-   ```
-   Task(
-     subagent_type="executor",
-     model="{executor_model}",
-     isolation="worktree",
-     team_name="maxsim-phase-{PHASE_NUMBER}-wave-{WAVE_NUM}",
-     prompt="
-       <objective>
-       Execute plan {plan_number} of phase {phase_number}-{phase_name}.
-       Commit each task atomically. Post SUMMARY as GitHub comment.
-       Move task sub-issues on the board: In Progress when started, Done when completed.
-       IMPORTANT: Do NOT update STATE.md or ROADMAP.md — the orchestrator handles metadata.
-       </objective>
-
-       <execution_context>
-       @./workflows/execute-plan.md
-       @./templates/summary.md
-       @./references/checkpoints.md
-       </execution_context>
-
-       <github_context>
-       Phase issue number: {phase_issue_number}
-       Plan comment body: {plan_comment_body}
-       Task mappings: {task_mappings_for_this_plan}
-       </github_context>
-
-       <files_to_read>
-       Read these files at execution start using the Read tool:
-       - .planning/STATE.md (State — READ ONLY, do not modify)
-       - .planning/config.json (Config, if exists)
-       - ./CLAUDE.md (Project instructions, if exists)
-       </files_to_read>
-
-       <constraints>
-       - You are running in a worktree. Do NOT modify .planning/STATE.md or .planning/ROADMAP.md.
-       - Only post summary comment to GitHub and commit task code.
-       - The orchestrator will handle metadata updates after all agents complete.
-       </constraints>
-
-       <success_criteria>
-       - [ ] All tasks executed
-       - [ ] Each task committed individually
-       - [ ] Summary posted as GitHub comment (type=summary) on phase issue #{phase_issue_number}
-       - [ ] Task sub-issues moved: In Progress → Done per task
-       </success_criteria>
-     ",
-     run_in_background=true
-   )
-   ```
-
-4. **Track progress with status table** (update as agents complete):
-   ```
-   | Plan | Agent | Status | Duration |
-   |------|-------|--------|----------|
-   | 05-01 | agent-abc | Running... | 2m |
-   | 05-02 | agent-def | Complete | 5m |
-   ```
-
-5. **After all agents in wave complete, collect results:**
-   - Check for `<!-- maxsim:type=summary -->` comments on the phase issue for each plan
-   - Run spot-checks (same as standard path: summary comment exists, commits present, sub-issues closed, review cycle check)
-   - Orchestrator performs batch metadata updates:
-     ```bash
-     for PLAN_ID in $COMPLETED_PLANS; do
-       node ~/.claude/maxsim/bin/maxsim-tools.cjs state advance-plan
-       node ~/.claude/maxsim/bin/maxsim-tools.cjs state update-progress
-     done
-     node ~/.claude/maxsim/bin/maxsim-tools.cjs roadmap update-plan-progress "${PHASE_NUMBER}"
-     ```
-   - Report:
-     ```
-     ---
-     ## Wave {N} Complete (Batch)
-
-     **{Plan ID}: {Plan Name}**
-     {What was built — from summary comment}
-     {Notable deviations, if any}
-
-     {If more waves: what this enables for next wave}
-     ---
-     ```
-
-6. **Cleanup worktrees after wave completes:**
-   ```bash
-   node ~/.claude/maxsim/bin/maxsim-tools.cjs worktree-cleanup --all
-   ```
-
-7. **Prepare inter-wave handoff context (for waves after Wave 1):**
-
-   When spawning agents for the next wave, include a brief context block so they can reference prior wave outputs without fetching full summary comments:
-
-   ```
-   <prior_wave_results>
-   Wave {N-1} completed:
-   {For each plan in prior wave: plan ID, one-liner from summary comment, key files created/modified}
-   </prior_wave_results>
-   ```
-
-   Add this block to the `<objective>` section of the next wave's Task() prompts.
-
-8. **Proceed to next wave.**
-
-**STANDARD PATH (when EXECUTION_MODE == "standard"):**
-
-**For each wave:**
-
-1. **Describe what's being built (BEFORE spawning):**
-
-   Read each plan's `<objective>` from the plan comment content. Extract what's being built and why.
-
-   ```
-   ---
-   ## Wave {N}
-
-   **{Plan ID}: {Plan Name}**
-   {2-3 sentences: what this builds, technical approach, why it matters}
-
-   Spawning {count} agent(s)...
-   ---
-   ```
-
-   - Bad: "Executing terrain generation plan"
-   - Good: "Procedural terrain generator using Perlin noise — creates height maps, biome zones, and collision meshes. Required before vehicle physics can interact with ground."
-
-2. **Spawn executor agents:**
-
-   Pass plan content from GitHub and GitHub context — executors do NOT need to read local PLAN.md files.
-   This keeps orchestrator context lean (~10-15%).
-
-   When the wave has multiple plans (parallel execution), add `team_name` to group agents:
-
-   ```
-   Task(
-     subagent_type="executor",
-     model="{executor_model}",
-     team_name="maxsim-phase-{PHASE_NUMBER}-wave-{WAVE_NUM}",  // only for multi-plan waves
-     prompt="
-       <objective>
-       Execute plan {plan_number} of phase {phase_number}-{phase_name}.
-       Commit each task atomically. Post SUMMARY as GitHub comment. Update STATE.md and ROADMAP.md.
-       Move task sub-issues on the board: In Progress when started, Done when completed.
-       </objective>
-
-       <execution_context>
-       @./workflows/execute-plan.md
-       @./templates/summary.md
-       @./references/checkpoints.md
-       @./references/tdd.md
-       </execution_context>
-
-       <github_context>
-       Phase issue number: {phase_issue_number}
-       Plan comment body: {plan_comment_body}
-       Task mappings: {task_mappings_for_this_plan}
-       </github_context>
-
-       <files_to_read>
-       Read these files at execution start using the Read tool:
-       - .planning/STATE.md (State)
-       - .planning/config.json (Config, if exists)
-       - ./CLAUDE.md (Project instructions, if exists — follow project-specific guidelines and coding conventions)
-       - .skills/ (Project skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)
-       </files_to_read>
-
-       <success_criteria>
-       - [ ] All tasks executed
-       - [ ] Each task committed individually
-       - [ ] Summary posted as GitHub comment (type=summary) on phase issue #{phase_issue_number}
-       - [ ] Task sub-issues moved: In Progress when started, Done when completed
-       - [ ] STATE.md updated with position and decisions
-       - [ ] ROADMAP.md updated with plan progress (via `roadmap update-plan-progress`)
-       </success_criteria>
-     "
-   )
-   ```
-
-3. **Wait for all agents in wave to complete.**
-
-4. **Report completion — spot-check claims and review cycle:**
-
-   For each completed plan:
-   - Fetch the phase issue comments and verify a `<!-- maxsim:type=summary -->` comment exists for this plan
-   - Check `git log --oneline --all --grep="{phase}-{plan}"` returns ≥1 commit
-   - Verify all task sub-issues for this plan are closed:
-     ```
-     mcp_list_sub_issues(issue_number: phase_issue_number)
-     ```
-   - Check for `## Self-Check: FAILED` in the summary comment body
-   - **Check for `## Review Cycle` section** in the summary comment — verify both review stages (Spec and Code) show PASS or SKIPPED (not BLOCKED or FAIL)
-
-   If ANY spot-check fails: report which plan failed, route to failure handler — ask "Retry plan?" or "Continue with remaining waves?"
-
-   If review cycle is missing or has unresolved issues: flag the plan as **review-incomplete** — ask "Run review cycle for this plan?" or "Continue (review will block phase completion)?"
-
-   **Note:** The executor agent runs the two-stage review (Spec Review + Code Review) after each wave. The orchestrator does NOT run reviews itself — it only checks the executor's review results in the summary comment. If review is missing, the executor failed to run it, and the orchestrator should offer to re-run the affected plan.
-
-   Review stages to check: `Spec:` and `Code:` lines in `## Review Cycle`. Both must be PASS or SKIPPED for the plan to be considered review-complete.
-
-   If pass — report:
-   ```
-   ---
-   ## Wave {N} Complete
-
-   **{Plan ID}: {Plan Name}**
-   {What was built — from summary comment}
-   {Notable deviations, if any}
-
-   {If more waves: what this enables for next wave}
-   ---
-   ```
-
-   - Bad: "Wave 2 complete. Proceeding to Wave 3."
-   - Good: "Terrain system complete — 3 biome types, height-based texturing, physics collision meshes. Vehicle physics (Wave 3) can now reference ground surfaces."
-
-5. **Handle failures:**
-
-   **Known Claude Code bug (classifyHandoffIfNeeded):** If an agent reports "failed" with error containing `classifyHandoffIfNeeded is not defined`, this is a Claude Code runtime bug — not a MAXSIM or agent issue. The error fires in the completion handler AFTER all tool calls finish. In this case: run the same spot-checks as step 4 (summary comment exists, git commits present, sub-issues closed, no Self-Check: FAILED). If spot-checks PASS → treat as **successful**. If spot-checks FAIL → treat as real failure below.
-
-   For real failures: report which plan failed → ask "Continue?" or "Stop?" → if continue, dependent plans may also fail. If stop, partial completion report.
-
-6. **Execute checkpoint plans between waves** — see `<checkpoint_handling>`.
-
-7. **Prepare inter-wave handoff context (for waves after Wave 1):**
-
-   When spawning agents for the next wave, include a brief context block so they can reference prior wave outputs without fetching full summary comments:
-
-   ```
-   <prior_wave_results>
-   Wave {N-1} completed:
-   {For each plan in prior wave: plan ID, one-liner from summary comment, key files created/modified}
-   </prior_wave_results>
-   ```
-
-   Add this block to the `<objective>` section of the next wave's Task() prompts.
-
-8. **Proceed to next wave.**
-</step>
-
-<step name="checkpoint_handling">
-Plans with `autonomous: false` require user interaction.
-
-**Auto-mode checkpoint handling:**
-
-Read auto-advance config:
-```bash
-AUTO_CFG=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs config-get workflow.auto_advance 2>/dev/null || echo "false")
-```
-
-When executor returns a checkpoint AND `AUTO_CFG` is `"true"`:
-- **human-verify** → Auto-spawn continuation agent with `{user_response}` = `"approved"`. Log `⚡ Auto-approved checkpoint`.
-- **decision** → Auto-spawn continuation agent with `{user_response}` = first option from checkpoint details. Log `⚡ Auto-selected: [option]`.
-- **human-action** → Present to user (existing behavior below). Auth gates cannot be automated.
-
-**Standard flow (not auto-mode, or human-action type):**
-
-1. Spawn agent for checkpoint plan
-2. Agent runs until checkpoint task or auth gate → returns structured state
-3. Agent return includes: completed tasks table, current task + blocker, checkpoint type/details, what's awaited
-4. **Present to user:**
-   ```
-   ## Checkpoint: [Type]
-
-   **Plan:** 03-03 Dashboard Layout
-   **Progress:** 2/3 tasks complete
-
-   [Checkpoint Details from agent return]
-   [Awaiting section from agent return]
-   ```
-5. User responds: "approved"/"done" | issue description | decision selection
-6. **Spawn continuation agent (NOT resume)** using continuation-prompt.md template:
-   - `{completed_tasks_table}`: From checkpoint return
-   - `{resume_task_number}` + `{resume_task_name}`: Current task
-   - `{user_response}`: What user provided
-   - `{resume_instructions}`: Based on checkpoint type
-   - `{github_context}`: Pass phase_issue_number and task_mappings for board transitions
-7. Continuation agent verifies previous commits, continues from resume point
-8. Repeat until plan completes or user stops
-
-**Why fresh agent, not resume:** Resume relies on internal serialization that breaks with parallel tool calls. Fresh agents with explicit state are more reliable.
-
-**Checkpoints in parallel waves:** Agent pauses and returns while other parallel agents may complete. Present checkpoint, spawn continuation, wait for all before next wave.
-</step>
-
-<step name="aggregate_results">
-After all waves:
-
-```markdown
-## Phase {X}: {Name} Execution Complete
-
-**Execution Mode:** {Batch (worktree) | Standard}
-**Worktrees Used:** {N} (batch only — omit for standard mode)
-**Waves:** {N} | **Plans:** {M}/{total} complete
-
-| Wave | Plans | Status |
-|------|-------|--------|
-| 1 | plan-01, plan-02 | ✓ Complete |
-| CP | plan-03 | ✓ Verified |
-| 2 | plan-04 | ✓ Complete |
-
-### Plan Details
-1. **03-01**: [one-liner from summary comment]
-2. **03-02**: [one-liner from summary comment]
-
-### Review Cycle Summary
-| Plan | Spec Review | Code Review | Retries |
-|------|-------------|-------------|---------|
-| 03-01 | PASS | PASS | 0 |
-| 03-02 | PASS | PASS | 1 |
-
-[Aggregate review findings from each plan's summary comment `## Review Cycle` section.
-If any plan has no Review Cycle section in its summary comment: mark as "NOT RUN" and flag for attention.
-If any plan has unresolved BLOCKED/FAIL status: list the blocking issues below.]
-
-### Unresolved Review Issues
-[List any plans with BLOCKED or FAIL review stages. These MUST be resolved before phase completion.]
-
-### Issues Encountered
-[Aggregate from summary comments, or "None"]
-```
-
-**Move phase issue to "In Review" on GitHub (WIRE-08):**
-All tasks are complete — move the phase to "In Review" before running verification:
-```
-mcp_move_issue(issue_number: phase_issue_number, status: "In Review")
-```
-
-**Phase completion gate:** If any plan has unresolved review issues (BLOCKED or FAIL in Spec Review or Code Review stages), the phase CANNOT proceed to `verify_phase_goal`. Present unresolved issues and offer:
-- "Fix review issues now" — re-run the review cycle for affected plans
-- "Override and continue" — mark as acknowledged, proceed (adds warning to VERIFICATION.md)
-</step>
-
-<step name="close_parent_artifacts">
-**For decimal/polish phases only (X.Y pattern):** Close the feedback loop by resolving parent UAT and debug artifacts.
-
-**Skip if** phase number has no decimal (e.g., `3`, `04`) — only applies to gap-closure phases like `4.1`, `03.1`.
-
-**1. Detect decimal phase and derive parent:**
-```bash
-# Check if phase_number contains a decimal
-if [[ "$PHASE_NUMBER" == *.* ]]; then
-  PARENT_PHASE="${PHASE_NUMBER%%.*}"
-fi
-```
-
-**2. Find parent UAT:**
-
-When GitHub integration is active, look for `<!-- maxsim:type=uat -->` comments on the parent phase issue:
-```
-mcp_get_issue_detail(issue_number: parent_phase_issue_number)
-```
-
-**If no parent UAT found:** Skip this step (gap-closure may have been triggered by VERIFICATION.md instead).
-
-**3. Update UAT gap statuses:**
-
-Read the parent UAT content (from comment or local file). For each gap entry with `status: failed`:
-- Update to `status: resolved`
-
-**4. If all gaps resolved — update UAT status:**
-
-- If posting to GitHub: update the UAT comment body with resolved statuses
-- If local file: update frontmatter `status: diagnosed` → `status: resolved`, update `updated:` timestamp
-
-**5. Resolve referenced debug sessions:**
-
-For each gap that has a `debug_session:` field:
-- Read the debug session file
-- Update frontmatter `status:` → `resolved`
-- Update frontmatter `updated:` timestamp
-- Move to resolved directory:
-```bash
-mkdir -p .planning/debug/resolved
-mv .planning/debug/{slug}.md .planning/debug/resolved/
-```
-
-**6. Commit updated artifacts:**
-```bash
-node ~/.claude/maxsim/bin/maxsim-tools.cjs commit "docs(phase-${PARENT_PHASE}): resolve UAT gaps and debug sessions after ${PHASE_NUMBER} gap closure" --files .planning/phases/*${PARENT_PHASE}*/*-UAT.md .planning/debug/resolved/*.md
-```
-</step>
-
-<step name="verify_phase_goal">
-Verify phase achieved its GOAL, not just completed tasks.
-
-```
-Task(
-  prompt="Verify phase {phase_number} goal achievement.
-Phase directory: {phase_dir}
-Phase issue: #{phase_issue_number}
-Phase goal: {goal from ROADMAP.md}
-Phase requirement IDs: {phase_req_ids}
-Check must_haves against actual codebase.
-Cross-reference requirement IDs from plan frontmatter against REQUIREMENTS.md — every ID MUST be accounted for.
-Post verification results as a GitHub comment (mcp_post_comment with type=verification) on phase issue #{phase_issue_number}.
-Also write VERIFICATION.md to the phase directory for local reference.",
-  subagent_type="verifier",
-  model="{verifier_model}"
-)
-```
-
-Read status from the verification comment posted to GitHub:
-```
-mcp_get_issue_detail(issue_number: phase_issue_number)
-```
-
-Look for `<!-- maxsim:type=verification -->` comment and parse `status:` field.
-
-| Status | Action |
-|--------|--------|
-| `passed` | → update_roadmap |
-| `human_needed` | Present items for human testing, get approval or feedback |
-| `gaps_found` | Present gap summary, offer `/maxsim:plan {phase} --gaps` |
-
-**If human_needed:**
-```
-## Phase {X}: {Name} — Human Verification Required
-
-All automated checks passed. {N} items need human testing:
-
-{From verification comment human_verification section}
-
-"approved" → continue | Report issues → gap closure
-```
-
-**If gaps_found:**
-```
-## Phase {X}: {Name} — Gaps Found
-
-**Score:** {N}/{M} must-haves verified
-**Report:** Phase issue #{phase_issue_number} (verification comment)
-
-### What's Missing
-{Gap summaries from verification comment}
-
----
-## Next Up
-
-`/maxsim:plan {X} --gaps`
-
-<sub>`/clear` first → fresh context window</sub>
-
-Also: view the verification comment on issue #{phase_issue_number} for the full report
-Also: `/maxsim:execute {X}` (includes verification) — manual testing first
-```
-
-Gap closure cycle: `/maxsim:plan {X} --gaps` reads verification comment → creates gap plans as new comments on phase issue (with `gap_closure: true` frontmatter) → user runs `/maxsim:execute {X} --gaps-only` → verifier re-runs.
-</step>
-
-<step name="update_roadmap">
-**Mark phase complete and update all tracking files:**
-
-```bash
-COMPLETION=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs phase complete "${PHASE_NUMBER}")
-```
-
-The CLI handles:
-- Marking phase checkbox `[x]` with completion date
-- Updating Progress table (Status → Complete, date)
-- Updating plan count to final
-- Advancing STATE.md to next phase
-- Updating REQUIREMENTS.md traceability
-
-Extract from result: `next_phase`, `next_phase_name`, `is_last_phase`.
-
-```bash
-node ~/.claude/maxsim/bin/maxsim-tools.cjs commit "docs(phase-{X}): complete phase execution" --files .planning/ROADMAP.md .planning/STATE.md .planning/REQUIREMENTS.md {phase_dir}/*-VERIFICATION.md
-```
-
-**Move phase issue to Done on GitHub:**
-```
-mcp_move_issue(issue_number: phase_issue_number, status: "Done")
-```
-
-**Post phase completion summary comment on the phase issue:**
-```
-mcp_post_comment(
-  issue_number: phase_issue_number,
-  type: "phase-complete",
-  body: "## Phase {phase_number} Execution Complete\n\nAll plans executed and verified.\n\n**Execution Mode:** {Batch | Standard}\n**Waves:** {wave_count}\n**Plans:** {completed}/{total}\n**Verification:** Passed\n\n### Plan Summary\n{one-liner per plan from summary comments}"
-)
-```
-
-</step>
-
-<step name="offer_next">
-
-**Exception:** If `gaps_found`, the `verify_phase_goal` step already presents the gap-closure path (`/maxsim:plan {X} --gaps`). No additional routing needed — skip auto-advance.
-
-**No-transition check (spawned by auto-advance chain):**
-
-Parse `--no-transition` flag from $ARGUMENTS.
-
-**If `--no-transition` flag present:**
-
-Execute-phase was spawned by plan's auto-advance. After verification passes and roadmap is updated, return completion status to parent:
-
-```
-## PHASE COMPLETE
-
-Phase: ${PHASE_NUMBER} - ${PHASE_NAME}
-Plans: ${completed_count}/${total_count}
-Verification: {Passed | Gaps Found}
-GitHub: Phase issue #{phase_issue_number} moved to Done
-
-[Include aggregate_results output]
-```
-
-STOP. Do not proceed to auto-advance or transition.
-
-**If `--no-transition` flag is NOT present:**
-
-**Auto-advance detection:**
-
-1. Parse `--auto` flag from $ARGUMENTS
-2. Read `workflow.auto_advance` from config:
-   ```bash
-   AUTO_CFG=$(node ~/.claude/maxsim/bin/maxsim-tools.cjs config-get workflow.auto_advance 2>/dev/null || echo "false")
-   ```
-
-**If `--auto` flag present OR `AUTO_CFG` is true (AND verification passed with no gaps):**
-
-```
-╔══════════════════════════════════════════╗
-║  AUTO-ADVANCING → TRANSITION             ║
-║  Phase {X} verified, continuing chain    ║
-╚══════════════════════════════════════════╝
-```
-
-Phase advancement is complete. The user may run `/maxsim:go` or `/maxsim:execute` to continue to the next phase.
-
-**If neither `--auto` nor `AUTO_CFG` is true:**
-
-The workflow ends. The user runs `/maxsim:go` or `/maxsim:execute` to advance to the next phase.
-</step>
-
-</process>
-
-<context_efficiency>
-Orchestrator: ~10-15% context. Subagents: fresh 200k each. No polling (Task blocks). No context bleed.
-Plan content passed directly from GitHub comments to subagents — no local PLAN.md reads in the orchestrator.
-</context_efficiency>
-
-<failure_handling>
-- **classifyHandoffIfNeeded false failure:** Agent reports "failed" but error is `classifyHandoffIfNeeded is not defined` → Claude Code bug, not MAXSIM. Spot-check (summary comment exists, commits present, sub-issues closed) → if pass, treat as success
-- **Agent fails mid-plan:** Missing summary comment on GitHub → report, ask user how to proceed
-- **Dependency chain breaks:** Wave 1 fails → Wave 2 dependents likely fail → user chooses attempt or skip
-- **All agents in wave fail:** Systemic issue → stop, report for investigation
-- **Checkpoint unresolvable:** "Skip this plan?" or "Abort phase execution?" → record partial progress in STATE.md
-- **GitHub integration unavailable:** Report error — GitHub Issues is the source of truth for plans and summaries
-</failure_handling>
-
-<resumption>
-Re-run `/maxsim:execute {phase}` → discover_and_group_plans fetches the phase issue comments and checks sub-issue closure → skips plans with all sub-issues closed → resumes from first incomplete plan → continues wave execution.
-
-STATE.md tracks: last completed plan, current wave, pending checkpoints.
-</resumption>