mapify-cli 3.6.0__tar.gz → 3.7.0__tar.gz

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/.gitignore

@@ -42,11 +42,13 @@ coverage.json
 .claude/sessions/
 .claude/metrics/
 .claude/cache/
+.claude/worktrees/
 
 # MAP runtime files (but keep static-analysis and orchestrator scripts)
 .map/*
 !.map/static-analysis/
 !.map/scripts/
+.map/scripts/.map/
 
 # Temporary verification files
 mapify_cli_verification_*.json

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mapify-cli
-Version: 3.6.0
+Version: 3.7.0
 Summary: MAP Framework installer - Modular Agentic Planner for Claude Code
 Project-URL: Homepage, https://github.com/azalio/map-framework
 Project-URL: Repository, https://github.com/azalio/map-framework.git
@@ -80,7 +80,6 @@ claude
 | `/map-efficient` | Production features, refactoring, complex tasks (recommended) |
 | `/map-debug` | Bug fixes and debugging |
 | `/map-fast` | Small, low-risk changes |
-| `/map-debate` | Complex decisions with multi-variant synthesis |
 | `/map-review` | Pre-commit code review |
 | `/map-check` | Quality gates and verification |
 | `/map-plan` | Task decomposition without implementation |
@@ -127,6 +126,10 @@ The orchestration lives in `.claude/commands/map-*.md` prompts created by `mapif
 | [Architecture](docs/ARCHITECTURE.md) | Agents, MCP integration, customization |
 | [Platform Spec](docs/MAP_PLATFORM_SPEC.md) | Platform refactor roadmap, codebase analysis |
 
+## Case Study
+
+- [DevOpsConf 2026 case study](https://github.com/azalio/devopsconf-ai-develop) — real production case: SPEC -> PLAN -> TEST -> CODE -> REVIEW -> LEARN
+
 ## Trouble?
 
 - **Command not found** → Run `mapify init` in your project first

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/README.md

@@ -48,7 +48,6 @@ claude
 | `/map-efficient` | Production features, refactoring, complex tasks (recommended) |
 | `/map-debug` | Bug fixes and debugging |
 | `/map-fast` | Small, low-risk changes |
-| `/map-debate` | Complex decisions with multi-variant synthesis |
 | `/map-review` | Pre-commit code review |
 | `/map-check` | Quality gates and verification |
 | `/map-plan` | Task decomposition without implementation |
@@ -95,6 +94,10 @@ The orchestration lives in `.claude/commands/map-*.md` prompts created by `mapif
 | [Architecture](docs/ARCHITECTURE.md) | Agents, MCP integration, customization |
 | [Platform Spec](docs/MAP_PLATFORM_SPEC.md) | Platform refactor roadmap, codebase analysis |
 
+## Case Study
+
+- [DevOpsConf 2026 case study](https://github.com/azalio/devopsconf-ai-develop) — real production case: SPEC -> PLAN -> TEST -> CODE -> REVIEW -> LEARN
+
 ## Trouble?
 
 - **Command not found** → Run `mapify init` in your project first

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "mapify-cli"
-version = "3.6.0"
+version = "3.7.0"
 description = "MAP Framework installer - Modular Agentic Planner for Claude Code"
 authors = [{ name = "MAP Framework Contributors" }]
 readme = "README.md"

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/__init__.py

@@ -23,7 +23,7 @@ Or install globally:
     mapify check
 """
 
-__version__ = "3.6.0"
+__version__ = "3.7.0"
 
 import os
 import subprocess

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/delivery/file_copier.py

@@ -312,7 +312,6 @@ This directory contains custom slash commands for Claude Code.
 - `/map-task` - Execute a single subtask from an existing plan
 - `/map-tdd` - Run a test-first workflow for one task or plan
 - `/map-debug` - Debug issues using MAP analysis
-- `/map-debate` - Generate variants and synthesize the best result
 - `/map-review` - Run a structured review workflow
 - `/map-check` - Run workflow quality gates and verification
 - `/map-fast` - Quick implementation with minimal validation

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/delivery/managed_file_copier.py

@@ -103,6 +103,17 @@ def inject_metadata(content: str, ext: str, version: str, template_hash: str) ->
 
     if ext == ".md":
         header = f"<!-- {_MANAGED_TAG}: {meta_json} -->\n"
+        # If content has YAML frontmatter (starts with ---), insert after
+        # closing --- to preserve frontmatter parsing by tools like Claude Code.
+        if content.startswith("---\n"):
+            end_idx = content.find("\n---\n", 3)
+            if end_idx != -1:
+                insert_pos = end_idx + 5  # after \n---\n
+                return content[:insert_pos] + header + content[insert_pos:]
+            # Edge case: closing --- at very end (no trailing newline)
+            end_idx = content.find("\n---", 3)
+            if end_idx != -1 and end_idx + 4 == len(content):
+                return content + "\n" + header
         return header + content
 
     if ext == ".py":
@@ -135,6 +146,7 @@ def extract_metadata(content: str, ext: str) -> tuple[Optional[dict[str, Any]],
     Returns (None, original_content) if no metadata found.
     """
     if ext == ".md":
+        # Try at start of file (non-frontmatter .md files)
         m = _MD_PATTERN.match(content)
         if m:
             try:
@@ -142,6 +154,27 @@ def extract_metadata(content: str, ext: str) -> tuple[Optional[dict[str, Any]],
                 return meta, content[m.end() :]
             except json.JSONDecodeError:
                 pass
+        # Try after YAML frontmatter (agent .md files with ---)
+        if content.startswith("---\n"):
+            end_idx = content.find("\n---\n", 3)
+            if end_idx != -1:
+                after_fm = end_idx + 5
+                rest = content[after_fm:]
+                m = _MD_PATTERN.match(rest)
+                if m:
+                    try:
+                        meta = json.loads(m.group(1))
+                        clean = content[:after_fm] + rest[m.end() :]
+                        # If nothing followed the MAP-MANAGED comment (it was
+                        # at EOF) and clean ends with "\n---\n", the trailing
+                        # newline was injected by inject_metadata for the
+                        # frontmatter-at-EOF edge case.  Strip it to restore
+                        # the original content that had no trailing newline.
+                        if not rest[m.end() :] and clean.endswith("\n---\n"):
+                            clean = clean[:-1]
+                        return meta, clean
+                    except json.JSONDecodeError:
+                        pass
         return None, content
 
     if ext == ".py":

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/repo_insight.py

@@ -7,6 +7,7 @@ and key directories for workflow initialization.
 from pathlib import Path
 from typing import List
 import json
+import subprocess
 
 
 def detect_language(project_root: Path) -> str:
@@ -201,3 +202,78 @@ def _validate_repo_insight_schema(data: dict) -> None:
     for dir_path in data["key_dirs"]:
         if dir_path.startswith("/"):
             raise ValueError(f"key_dirs must be relative paths: {dir_path}")
+
+
+def compute_differential_insight(project_root: Path, since_sha: str | None) -> dict:
+    """Compute file changes since a given git SHA.
+
+    Used for context-aware injection: shows Actor only files
+    that changed since the last subtask completed.
+
+    Args:
+        project_root: Path to project root
+        since_sha: Git SHA to diff against (None = no baseline)
+
+    Returns:
+        Dict with changed_files, deleted_files. On success also includes
+        since_sha and current_sha. On error: empty lists and error key.
+        When since_sha is None: empty lists and note key.
+    """
+    if since_sha is None:
+        return {"changed_files": [], "deleted_files": [], "note": "no baseline SHA"}
+
+    try:
+        # Get changed/added/modified/renamed files
+        result = subprocess.run(
+            ["git", "diff", "--name-only", "--diff-filter=ACMR", since_sha, "HEAD"],
+            capture_output=True,
+            text=True,
+            cwd=project_root,
+            timeout=2,
+        )
+        if result.returncode != 0:
+            return {
+                "changed_files": [],
+                "deleted_files": [],
+                "error": f"git diff failed: {result.stderr.strip()}",
+            }
+        changed = [f for f in result.stdout.strip().split("\n") if f]
+
+        # Get deleted files
+        result_del = subprocess.run(
+            ["git", "diff", "--name-only", "--diff-filter=D", since_sha, "HEAD"],
+            capture_output=True,
+            text=True,
+            cwd=project_root,
+            timeout=2,
+        )
+        deleted = (
+            [f for f in result_del.stdout.strip().split("\n") if f]
+            if result_del.returncode == 0
+            else []
+        )
+
+        # Get current HEAD SHA
+        head_result = subprocess.run(
+            ["git", "rev-parse", "HEAD"],
+            capture_output=True,
+            text=True,
+            cwd=project_root,
+            timeout=2,
+        )
+        current_sha = (
+            head_result.stdout.strip() if head_result.returncode == 0 else "unknown"
+        )
+
+        return {
+            "changed_files": changed,
+            "deleted_files": deleted,
+            "since_sha": since_sha,
+            "current_sha": current_sha,
+        }
+    except (subprocess.TimeoutExpired, FileNotFoundError, OSError) as e:
+        return {
+            "changed_files": [],
+            "deleted_files": [],
+            "error": str(e),
+        }

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/templates/agents/actor.md

@@ -200,10 +200,15 @@ Task(
    - Use Read(path, offset=lines[0], limit=lines[1]-lines[0]+1)  # lines = [start, end], inclusive
    - Don't read all locations — only what you actually need
 
-## Skip Research If
+## Research Usage
 
-- Task is self-contained (new file, no dependencies)
-- Existing patterns from context already cover the need
+Research is run by the orchestrator BEFORE Actor is invoked. The findings file
+(`.map/<branch>/findings_<branch>.md`) contains distilled context. If it exists,
+read it before implementation — it has import patterns, module structure, and
+build configuration that prevent integration failures.
+
+Do NOT skip reading the findings file even for "new file" tasks — new files still
+need correct imports, types, and build configuration from the existing project.
 
 ---
 
@@ -276,6 +281,9 @@ Rules:
 5. Include edge cases from the spec's `## Edge Cases` section if available in the packet.
 6. Use standard test patterns for the project's language and framework.
 7. Tests SHOULD fail when run (implementation doesn't exist yet). This is expected.
+8. Do NOT add temporal comments about test failure status (e.g., "currently FAILS",
+   "expected to FAIL", "will PASS once fix is applied"). Write tests as permanent,
+   clean code — the Red/Green state is transient and must not leak into comments.
 
 Output:
 - Test files created via Write tool
@@ -408,7 +416,17 @@ VC1: <criterion text>
 - Tests: path/to/test_file.ext::test_name (or N/A with reason)
 ```
 
-## 7. Integration Notes (If Applicable)
+## 7. Downstream Consumption Check
+
+When implementing a component whose output is consumed by another component:
+
+- **Identify the consumer**: What reads your output? Verify your output populates ALL fields it expects.
+- **Self-bootstrap**: Does your code load its own dependencies from config/storage, or does it silently return empty results when input is not pre-populated by the caller?
+- **Stub replacement**: If implementing a real version of a placeholder, verify it is wired into the runtime — not just available as a standalone function.
+
+Skip this section for leaf components with no downstream consumers.
+
+## 8. Integration Notes (If Applicable)
 
 Only include if changes affect:
 - Database schema (migrations needed?)

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/templates/agents/monitor.md

@@ -16,16 +16,17 @@ You are a Protocol-Driven Validation System. Your objective: verify that Actor's
 
 **CRITICAL: Monitor is READ-ONLY reviewer, NOT a code editor**
 
-You are a **validation agent**, NOT a code executor. Your role:
+You are a **validation agent**, NOT a code editor. Your role:
 
 - ✅ DO: Review Actor's code proposals and output JSON feedback
 - ✅ DO: Use Read tool to examine existing code for context
+- ✅ DO: Run read-only build/test commands (tsc --noEmit, go build, pytest, etc.) to verify code compiles and passes
 - ❌ NEVER: Use Edit or MultiEdit tools
 - ⚠️ EXCEPTION: Write tool is permitted ONLY for evidence artifacts (.map/ directory)
 - ❌ NEVER: Modify source files directly
 - ❌ NEVER: "Fix code for Actor" - only REPORT issues
 - 📋 WHY: workflow-gate.py will BLOCK Edit and non-evidence Write during monitor phase
-- 🔄 FLOW: Actor outputs → **You review** → Orchestrator applies (if approved)
+- 🔄 FLOW: Actor outputs → **You review + run build/tests** → Orchestrator applies (if approved)
 
 **Your output**: JSON with `valid: true|false` and `issues[]` array
 
@@ -39,10 +40,16 @@ You are a **validation agent**, NOT a code executor. Your role:
 
 **Verification sequence (execute in order):**
 1. Parse AAG contract from prompt — extract Actor, Action, Goal
-2. Verify Goal is achieved — trace code path to confirm the stated outcome
-3. Verify Action is implemented — check that the specified method/operation exists
-4. Verify scope — confirm changes stay within Actor's allowed_scope
-5. Run quality gates below
+2. **BUILD GATE (MANDATORY — run FIRST):** Run the project's build/compile command:
+   - TypeScript: `npx tsc --noEmit` (or `npm run build`)
+   - Python: `python -m py_compile <changed_files>` (or mypy if configured)
+   - Go: `go build ./...`
+   - Rust: `cargo check`
+   - If build/compile fails → `valid: false` immediately with compilation errors. Do NOT proceed to other checks.
+3. Verify Goal is achieved — trace code path to confirm the stated outcome
+4. Verify Action is implemented — check that the specified method/operation exists
+5. Verify scope — confirm changes stay within Actor's allowed_scope
+6. Run quality gates below
 
 **Deterministic REJECT rule:**
 If implementation deviates from the AAG contract — `valid: false` — regardless of how "clean" or "elegant" the code is. The contract IS the specification; aesthetic quality is irrelevant when the contract is violated.
@@ -50,15 +57,16 @@ If implementation deviates from the AAG contract — `valid: false` — regardle
 **Escalation Framework:**
 
 🔴 **AUTO-REJECT (valid: false, must fix):**
-1. **AAG contract violation** — implementation does not satisfy Actor -> Action -> Goal
-2. Missing error handling on network/database/file operations
-3. No input validation on user-provided data
-4. SQL string concatenation (injection vulnerability)
-5. Hardcoded secrets (API keys, passwords, tokens)
-6. Silent failures (try/catch with empty handler)
-7. Deprecated APIs without migration plan
-8. Security score < 7 OR functionality score < 7
-9. **Missing intent comments** — non-obvious logic blocks without `# Intent: <why>` comments, or removal of existing intent comments that describe author's reasoning
+1. **Build/compile failure** — code does not compile (`tsc --noEmit`, `go build`, `cargo check`, `py_compile` fails)
+2. **AAG contract violation** — implementation does not satisfy Actor -> Action -> Goal
+3. Missing error handling on network/database/file operations
+4. No input validation on user-provided data
+5. SQL string concatenation (injection vulnerability)
+6. Hardcoded secrets (API keys, passwords, tokens)
+7. Silent failures (try/catch with empty handler)
+8. Deprecated APIs without migration plan
+9. Security score < 7 OR functionality score < 7
+10. **Missing intent comments** — non-obvious logic blocks without `# Intent: <why>` comments, or removal of existing intent comments that describe author's reasoning
 
 🟡 **WARN (should address, not blocking):**
 1. Missing edge case tests (empty arrays, null values)
@@ -280,8 +288,8 @@ IF detected_language != "unknown":
   → Consider language-specific static analysis tools
 
 PHASE 3: EXHAUSTIVE DIMENSION VALIDATION (ALWAYS)
-Execute validation protocol for each of the 10 dimensions sequentially.
-Do NOT skip dimensions based on early findings — complete ALL 10.
+Execute validation protocol for each of the 11 dimensions sequentially.
+Do NOT skip dimensions based on early findings — complete ALL 11.
 For each dimension: parse criteria → verify against code → record PASS/FAIL.
 Apply language-specific validation rules per dimension.
 
@@ -852,11 +860,11 @@ Include in JSON output when validation_criteria provided:
 
 </Monitor_Contract_Validation>
 
-<Monitor_10D_Validation_v2_9>
+<Monitor_11D_Validation_v3_0>
 
-## 10-Dimension Quality Model
+## 11-Dimension Quality Model
 
-Execute validation protocol for EACH dimension sequentially. Do NOT short-circuit — complete ALL 10 dimensions even if early rejections found. Output structured findings per dimension.
+Execute validation protocol for EACH dimension sequentially. Do NOT short-circuit — complete ALL 11 dimensions even if early rejections found. Output structured findings per dimension. **Exception:** BUILD GATE failure (step 2 of Verification sequence) is the single allowed short-circuit — if build/compile fails, set `valid: false` immediately without completing dimension checks.
 
 ### 1. CORRECTNESS
 
@@ -1374,7 +1382,47 @@ ELSE:
 - Post-cutoff library + no research + outdated patterns
 </critical>
 
-</Monitor_10D_Validation_v2_9>
+### 11. INTEGRATION (When subtask has upstream/downstream dependencies)
+
+#### What to Check
+- Output consumed correctly by downstream components (not silently dropped)
+- Component self-bootstraps from config/storage (does not require caller to pre-populate dependencies)
+- Stubs/placeholders replaced by real implementations in the runtime entrypoint
+- Interface contracts between components are satisfied in both directions
+
+#### How to Check
+1. Identify downstream consumers of this subtask's output
+2. Trace the data path: does the output reach the consumer with the expected shape?
+3. Check if the component loads its own dependencies or silently returns empty/stub results
+4. Verify the runtime entrypoint uses the real implementation, not a placeholder
+
+#### Pass Criteria
+- Output is demonstrably consumed by at least one downstream component
+- Component works when invoked through the runtime entrypoint (not just direct calls)
+- No silent fallback to stub/empty results on missing dependencies
+
+#### Common Failures
+- Component writes to field A but consumer reads field B
+- Runtime entrypoint still wired to a stub despite real implementation existing
+- Component returns empty results when dependencies are not injected by test setup
+- Build/config failure masked as a successful stub response
+
+#### Severity Mapping
+- **Critical**: Runtime entrypoint returns stub/placeholder to end users
+- **High**: Component output not consumed by downstream (data silently lost)
+- **Medium**: Component requires caller injection instead of self-bootstrapping
+- **Low**: Interface contract undocumented but happens to work
+
+**Decision Framework**:
+```
+IF subtask has no downstream consumers AND no runtime entrypoint:
+  → Skip (leaf component)
+ELSE:
+  → Verify output reaches consumer through runtime path
+  → Verify self-bootstrapping from config/storage
+```
+
+</Monitor_11D_Validation_v3_0>
 
 
 <Monitor_Severity_Matrix>
@@ -1395,6 +1443,7 @@ This table provides quick reference for severity classification per dimension. U
 | **8. External Deps** | Missing critical dependency documentation | Incomplete CRD/adapter docs | Missing version constraints | Minor config details |
 | **9. Documentation** | Contradicts tech-design/source of truth | Missing key fields/logic, wrong ownership | Minor inconsistencies | Formatting issues |
 | **10. Research** | N/A (rarely critical) | Complex problem + no research + wrong impl | Post-cutoff lib + outdated patterns | Missing citations only |
+| **11. Integration** | Runtime entrypoint returns stub to users | Output not consumed by downstream (data lost) | Requires caller injection instead of self-bootstrap | Interface contract undocumented |
 
 ### Severity Decision Tree
 
@@ -1553,8 +1602,8 @@ Do NOT invent issues to justify review effort. Empty `issues` array is valid.
           },
           "category": {
             "type": "string",
-            "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research"],
-            "description": "Maps to 10-dimension model: 1=correctness, 2=security, 3=code-quality, 4=performance, 5=testability, 6=cli-tool, 7=maintainability, 8=external-deps, 9=documentation, 10=research"
+            "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research", "integration"],
+            "description": "Maps to 11-dimension model: 1=correctness, 2=security, 3=code-quality, 4=performance, 5=testability, 6=cli-tool, 7=maintainability, 8=external-deps, 9=documentation, 10=research, 11=integration"
           },
           "title": {
             "type": "string",
@@ -1601,7 +1650,7 @@ Do NOT invent issues to justify review effort. Empty `issues` array is valid.
       "type": "array",
       "items": {
         "type": "string",
-        "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research"]
+        "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research", "integration"]
       },
       "description": "Dimensions that passed completely"
     },
@@ -1609,7 +1658,7 @@ Do NOT invent issues to justify review effort. Empty `issues` array is valid.
       "type": "array",
       "items": {
         "type": "string",
-        "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research"]
+        "enum": ["correctness", "security", "code-quality", "performance", "testability", "cli-tool", "maintainability", "external-deps", "documentation", "research", "integration"]
       },
       "description": "Dimensions with issues"
     },
@@ -1908,7 +1957,7 @@ ARRAY POPULATION:
 - Ensure: passed_checks ∩ failed_checks = ∅ (no overlap)
 
 SPECIAL CASES:
-- If no issues found: all 10 categories go in passed_checks
+- If no issues found: all 11 categories go in passed_checks
 - If a dimension was skipped (large change): omit from both arrays
 ```
 
@@ -2016,6 +2065,7 @@ ELSE:
 | `external-deps` | Missing CRDs, undocumented dependencies | 8 |
 | `documentation` | Inconsistent with source, missing fields | 9 |
 | `research` | Missing research for unfamiliar patterns | 10 |
+| `integration` | Output not consumed downstream, stub in runtime | 11 |
 
 </Monitor_Decision_Rules>
 
@@ -2458,7 +2508,7 @@ def check_rate_limit(user_id, action, limit=100, window=3600):
 
 1. ✅ Did I use request_review for code implementations?
 2. ✅ Did I check for known issue patterns?
-3. ✅ Did I check all 10 validation dimensions systematically?
+3. ✅ Did I check all 11 validation dimensions systematically?
 4. ✅ Did I verify documentation against source of truth (if applicable)?
 5. ✅ Are all issues specific with location and actionable suggestions?
 6. ✅ Is severity classification correct per guidelines?

{mapify_cli-3.6.0 → mapify_cli-3.7.0}/src/mapify_cli/templates/agents/task-decomposer.md

@@ -278,12 +278,24 @@ Return **ONLY** valid JSON in this exact structure:
     - In that case: either add a FOUNDATION subtask to introduce a minimal test harness, or document the gap explicitly in risks/assumptions.
 **subtasks[].affected_files**: Precise file paths (NOT "backend", "frontend"); use [] if paths unknown
 
+### Integration & Runtime Bootstrapping Subtasks
+
+Feature subtasks implement components in isolation. To ensure they work together in the real runtime, you MUST also create:
+
+1. **Integration subtask** (one per runtime entrypoint): Wires real implementations into the runtime surface, replacing any stubs/placeholders. AAG contract must name the entrypoint and verify end-to-end data flow through it.
+   - Depends on ALL feature subtasks it integrates.
+
+2. **Bootstrapping subtask** (when components need external data at runtime): Ensures each workflow loads its own dependencies from configuration or persistent storage rather than requiring callers to pre-populate them.
+
+3. **Interface contracts between subtasks**: When subtask A produces output consumed by subtask B, document the data contract in BOTH subtasks' validation criteria so neither side can silently break it.
+
 ### Subtask Ordering
 
 Subtasks should be ordered by dependency:
 1. Foundation subtasks (no dependencies) first
 2. Dependent subtasks after their prerequisites
-3. Tests/docs can be parallel with implementation (same dependency level)
+3. Integration/wiring subtasks after ALL feature subtasks they integrate
+4. Tests/docs can be parallel with implementation (same dependency level)
 
 **CRITICAL**: If subtask B depends on subtask A, A must appear BEFORE B in the array.
 
@@ -529,6 +541,12 @@ When invoked with `mode: "re_decomposition"` from the orchestrator, you receive
 - [ ] Dependencies reference valid subtask IDs
 - [ ] Follows ordering constraint (dependencies before dependents)
 
+**Integration & Wiring**:
+- [ ] At least one integration subtask wires features into each runtime entrypoint
+- [ ] Interface contracts documented when one subtask produces output consumed by another
+- [ ] Bootstrapping subtask exists if components need data from disk/config at runtime
+- [ ] No subtask silently assumes its output is consumed — explicit consumer named in VC
+
 **Dependency Validation** (CRITICAL):
 - [ ] **Circular dependency check**: Verify dependency graph is acyclic (A→B→C→A is INVALID)
 - [ ] **Mental topological sort**: Can all subtasks be executed in a valid order?