icdev 1.0.0__py3-none-any.whl

icdev/tools/requirements/decomposition_engine.py

@@ -0,0 +1,778 @@
+#!/usr/bin/env python3
+# CUI // SP-CTI
+# Controlled by: Department of Defense
+# CUI Category: CTI
+# Distribution: D
+# POC: ICDEV System Administrator
+"""SAFe Agile hierarchy decomposition engine.
+
+Decomposes intake session requirements into SAFe hierarchy:
+Epic > Capability > Feature > Story > Enabler.
+
+Includes T-shirt sizing, WSJF scoring, BDD acceptance criteria generation,
+and ATO impact tier assignment.
+
+Usage:
+    # Decompose requirements to story level
+    python tools/requirements/decomposition_engine.py --session-id sess-abc --json
+
+    # Decompose with BDD criteria generation
+    python tools/requirements/decomposition_engine.py --session-id sess-abc \\
+        --level story --generate-bdd --json
+
+    # Get existing decomposition
+    python tools/requirements/decomposition_engine.py --session-id sess-abc --get --json
+
+    # Get existing decomposition filtered by level
+    python tools/requirements/decomposition_engine.py --session-id sess-abc \\
+        --get --level feature --json
+"""
+
+import argparse
+import json
+import os
+import re
+import sqlite3
+import uuid
+from datetime import datetime
+from pathlib import Path
+from icdev._paths import get_project_root
+
+BASE_DIR = get_project_root()
+DB_PATH = Path(os.environ.get("ICDEV_DB_PATH", str(BASE_DIR / "data" / "icdev.db")))
+
+try:
+    from icdev.tools.audit.audit_logger import log_event
+    _HAS_AUDIT = True
+except ImportError:
+    _HAS_AUDIT = False
+    def log_event(**kwargs): return -1
+
+
+def _get_connection(db_path=None):
+    """Get database connection with dict-like row access."""
+    path = db_path or DB_PATH
+    if not path.exists():
+        raise FileNotFoundError(
+            f"Database not found: {path}\nRun: python tools/db/init_icdev_db.py"
+        )
+    conn = sqlite3.connect(str(path))
+    conn.row_factory = sqlite3.Row
+    return conn
+
+
+def _generate_id(prefix="safe"):
+    """Generate a unique ID with prefix."""
+    return f"{prefix}-{uuid.uuid4().hex[:12]}"
+
+
+# ---------------------------------------------------------------------------
+# T-shirt sizing and WSJF scoring
+# ---------------------------------------------------------------------------
+
+_TSHIRT_NUMERIC = {"XS": 1, "S": 1, "M": 2, "L": 3, "XL": 5, "XXL": 8}
+
+_PRIORITY_BUSINESS_VALUE = {
+    "critical": 10,
+    "high": 7,
+    "medium": 4,
+    "low": 2,
+}
+
+# ATO impact tier keywords
+_ATO_TIER_KEYWORDS = {
+    "RED": [
+        "secret", "ts/sci", "top secret", "classified",
+        "authorization boundary", "ato boundary",
+    ],
+    "ORANGE": [
+        "external system", "new interface", "new integration",
+        "mobile", "byod", "cloud migration",
+        "fips", "encryption change", "pki",
+    ],
+    "YELLOW": [
+        "access control", "rbac", "audit", "logging",
+        "authentication", "cac", "mfa", "stig",
+    ],
+}
+
+
+def _estimate_tshirt(req_count):
+    """Estimate T-shirt size from requirement count.
+
+    S=1, M=2-3, L=4-5, XL=6+
+    """
+    if req_count <= 1:
+        return "S"
+    elif req_count <= 3:
+        return "M"
+    elif req_count <= 5:
+        return "L"
+    else:
+        return "XL"
+
+
+def _compute_wsjf(priority, t_shirt_size):
+    """Compute WSJF score: business_value / t_shirt_numeric.
+
+    business_value: 1-10 based on priority.
+    t_shirt_numeric: S=1, M=2, L=3, XL=5, XXL=8.
+    """
+    bv = _PRIORITY_BUSINESS_VALUE.get(priority, 4)
+    divisor = _TSHIRT_NUMERIC.get(t_shirt_size, 2)
+    if divisor == 0:
+        divisor = 1
+    return round(bv / divisor, 2)
+
+
+def _determine_ato_impact(text):
+    """Determine ATO impact tier from requirement text."""
+    lower = text.lower()
+    for tier in ("RED", "ORANGE", "YELLOW"):
+        keywords = _ATO_TIER_KEYWORDS[tier]
+        if any(kw in lower for kw in keywords):
+            return tier
+    return "GREEN"
+
+
+def _determine_dominant_priority(reqs):
+    """Get the highest priority from a list of requirements."""
+    priority_order = ["critical", "high", "medium", "low"]
+    for p in priority_order:
+        if any(r.get("priority") == p for r in reqs):
+            return p
+    return "medium"
+
+
+# ---------------------------------------------------------------------------
+# BDD generation
+# ---------------------------------------------------------------------------
+
+def generate_bdd_criteria(requirement_text, requirement_type):
+    """Generate Gherkin Given/When/Then from requirement text.
+
+    Pattern: Given [context from requirement], When [action implied],
+    Then [expected outcome].
+    """
+    text = requirement_text.strip()
+    # Truncate for scenario name
+    scenario_name = text[:60] + ("..." if len(text) > 60 else "")
+
+    # Build context from requirement type
+    type_contexts = {
+        "functional": "the system is operational and the user is authenticated",
+        "security": "the system enforces security controls per the accreditation boundary",
+        "performance": "the system is under normal operational load",
+        "interface": "all external system interfaces are connected and operational",
+        "data": "the data store is initialized and accessible",
+        "compliance": "the system is deployed within the authorized environment",
+        "operational": "the system is in its operational environment",
+        "non_functional": "the system meets baseline non-functional requirements",
+        "constraint": "all system constraints and limitations are documented",
+        "transitional": "the migration or transition plan is in effect",
+    }
+    given_context = type_contexts.get(requirement_type, "the system is operational")
+
+    # Extract action from text heuristics
+    # Look for verb phrases after shall/must/should/will
+    action_match = re.search(
+        r"(?:shall|must|should|will|needs?\s+to|is\s+required\s+to)\s+(.+)",
+        text,
+        re.IGNORECASE,
+    )
+    if action_match:
+        action_text = action_match.group(1).strip().rstrip(".")
+        # Capitalize first letter
+        action_text = action_text[0].upper() + action_text[1:] if action_text else text
+    else:
+        action_text = text[:80]
+
+    # Build the Gherkin
+    gherkin = (
+        f"Feature: {requirement_type.replace('_', ' ').title()} Requirement Validation\n"
+        f"\n"
+        f"  Scenario: {scenario_name}\n"
+        f"    Given {given_context}\n"
+        f"    When {action_text}\n"
+        f"    Then the system behaves as specified and the requirement is satisfied"
+    )
+
+    return gherkin
+
+
+# ---------------------------------------------------------------------------
+# Core decomposition
+# ---------------------------------------------------------------------------
+
+def decompose_requirements(
+    session_id,
+    target_level="story",
+    generate_bdd=False,
+    estimate=True,
+    db_path=None,
+):
+    """Decompose session requirements into SAFe hierarchy.
+
+    Reads all requirements from intake_requirements where session_id matches.
+    Groups by requirement_type into Epics, breaks each Epic into Features,
+    and for each Feature creates Stories (one per requirement) plus an
+    Enabler for infrastructure needs.
+
+    Args:
+        session_id: Intake session ID.
+        target_level: Decompose to this level (epic, feature, story).
+        generate_bdd: If True, generate Given/When/Then acceptance criteria.
+        estimate: If True, assign T-shirt sizes and WSJF scores.
+        db_path: Optional DB path override.
+
+    Returns:
+        dict with session_id, items_created count, level breakdown, and items list.
+    """
+    conn = _get_connection(db_path)
+
+    # Verify session exists
+    session = conn.execute(
+        "SELECT * FROM intake_sessions WHERE id = ?", (session_id,)
+    ).fetchone()
+    if not session:
+        conn.close()
+        raise ValueError(f"Session '{session_id}' not found.")
+
+    session_data = dict(session)
+    project_id = session_data.get("project_id")
+
+    # Load all non-rejected requirements for the session
+    reqs = conn.execute(
+        "SELECT * FROM intake_requirements WHERE session_id = ? AND status != 'rejected'",
+        (session_id,),
+    ).fetchall()
+    reqs = [dict(r) for r in reqs]
+
+    if not reqs:
+        conn.close()
+        return {
+            "status": "ok",
+            "session_id": session_id,
+            "items_created": 0,
+            "levels": {"epics": 0, "features": 0, "stories": 0, "enablers": 0},
+            "items": [],
+            "message": "No requirements to decompose.",
+        }
+
+    items = []
+    counts = {"epics": 0, "features": 0, "stories": 0, "enablers": 0}
+
+    # --- Group into Epics by requirement_type ---
+    type_groups = {}
+    for req in reqs:
+        rtype = req.get("requirement_type", "functional")
+        type_groups.setdefault(rtype, []).append(req)
+
+    for rtype, group_reqs in type_groups.items():
+        # --- Create Epic ---
+        epic_id = _generate_id("epic")
+        epic_title = f"{rtype.replace('_', ' ').title()} Capabilities"
+        epic_description = (
+            f"All {rtype} requirements for the system. "
+            f"Contains {len(group_reqs)} requirement(s)."
+        )
+        req_ids = [r["id"] for r in group_reqs]
+        epic_priority = _determine_dominant_priority(group_reqs)
+        epic_tshirt = _estimate_tshirt(len(group_reqs)) if estimate else None
+        epic_wsjf = _compute_wsjf(epic_priority, epic_tshirt) if estimate and epic_tshirt else None
+
+        # ATO impact from all requirement text
+        all_text = " ".join(r.get("raw_text", "") for r in group_reqs)
+        epic_ato = _determine_ato_impact(all_text)
+
+        epic_bdd = None
+        if generate_bdd:
+            epic_bdd = generate_bdd_criteria(epic_description, rtype)
+
+        conn.execute(
+            """INSERT INTO safe_decomposition
+               (id, session_id, project_id, parent_id, level, title,
+                description, acceptance_criteria, t_shirt_size, story_points,
+                pi_target, wsjf_score, ato_impact_tier, source_requirement_ids,
+                status, classification, created_at)
+               VALUES (?, ?, ?, NULL, 'epic', ?, ?, ?, ?, NULL, NULL, ?, ?, ?, 'draft', 'CUI', ?)""",
+            (
+                epic_id, session_id, project_id,
+                epic_title, epic_description, epic_bdd,
+                epic_tshirt, epic_wsjf, epic_ato,
+                json.dumps(req_ids),
+                datetime.now().isoformat(),
+            ),
+        )
+
+        epic_item = {
+            "id": epic_id,
+            "level": "epic",
+            "title": epic_title,
+            "description": epic_description,
+            "t_shirt_size": epic_tshirt,
+            "wsjf_score": epic_wsjf,
+            "ato_impact_tier": epic_ato,
+            "parent_id": None,
+            "children_count": len(group_reqs),
+            "status": "draft",
+        }
+        if generate_bdd:
+            epic_item["acceptance_criteria"] = epic_bdd
+        items.append(epic_item)
+        counts["epics"] += 1
+
+        if target_level in ("feature", "story", "enabler"):
+            # --- Create Features: one per requirement ---
+            for req in group_reqs:
+                feature_id = _generate_id("feat")
+                raw_text = req.get("raw_text", "")
+                feature_title = raw_text[:80] if len(raw_text) > 80 else raw_text
+                feature_description = raw_text
+                feature_priority = req.get("priority", "medium")
+                feature_tshirt = _estimate_tshirt(1) if estimate else None  # one req = S
+                feature_wsjf = (
+                    _compute_wsjf(feature_priority, feature_tshirt)
+                    if estimate and feature_tshirt else None
+                )
+                feature_ato = _determine_ato_impact(raw_text)
+
+                feature_bdd = None
+                if generate_bdd:
+                    feature_bdd = generate_bdd_criteria(raw_text, rtype)
+
+                conn.execute(
+                    """INSERT INTO safe_decomposition
+                       (id, session_id, project_id, parent_id, level, title,
+                        description, acceptance_criteria, t_shirt_size, story_points,
+                        pi_target, wsjf_score, ato_impact_tier, source_requirement_ids,
+                        status, classification, created_at)
+                       VALUES (?, ?, ?, ?, 'feature', ?, ?, ?, ?, NULL, NULL, ?, ?, ?, 'draft', 'CUI', ?)""",
+                    (
+                        feature_id, session_id, project_id, epic_id,
+                        feature_title, feature_description, feature_bdd,
+                        feature_tshirt, feature_wsjf, feature_ato,
+                        json.dumps([req["id"]]),
+                        datetime.now().isoformat(),
+                    ),
+                )
+
+                feature_item = {
+                    "id": feature_id,
+                    "level": "feature",
+                    "title": feature_title,
+                    "description": feature_description,
+                    "t_shirt_size": feature_tshirt,
+                    "wsjf_score": feature_wsjf,
+                    "ato_impact_tier": feature_ato,
+                    "parent_id": epic_id,
+                    "status": "draft",
+                }
+                if generate_bdd:
+                    feature_item["acceptance_criteria"] = feature_bdd
+                items.append(feature_item)
+                counts["features"] += 1
+
+                if target_level in ("story", "enabler"):
+                    # --- Create Story for this requirement ---
+                    story_id = _generate_id("story")
+                    story_title = raw_text[:80] if len(raw_text) > 80 else raw_text
+                    story_description = raw_text
+                    story_tshirt = "S" if estimate else None  # one req = S
+                    story_points = _TSHIRT_NUMERIC.get(story_tshirt, 1) if estimate else None
+                    story_wsjf = (
+                        _compute_wsjf(feature_priority, story_tshirt)
+                        if estimate and story_tshirt else None
+                    )
+                    story_ato = feature_ato
+
+                    story_bdd = None
+                    if generate_bdd:
+                        story_bdd = generate_bdd_criteria(raw_text, rtype)
+
+                    conn.execute(
+                        """INSERT INTO safe_decomposition
+                           (id, session_id, project_id, parent_id, level, title,
+                            description, acceptance_criteria, t_shirt_size, story_points,
+                            pi_target, wsjf_score, ato_impact_tier, source_requirement_ids,
+                            status, classification, created_at)
+                           VALUES (?, ?, ?, ?, 'story', ?, ?, ?, ?, ?, NULL, ?, ?, ?, 'draft', 'CUI', ?)""",
+                        (
+                            story_id, session_id, project_id, feature_id,
+                            story_title, story_description, story_bdd,
+                            story_tshirt, story_points, story_wsjf, story_ato,
+                            json.dumps([req["id"]]),
+                            datetime.now().isoformat(),
+                        ),
+                    )
+
+                    story_item = {
+                        "id": story_id,
+                        "level": "story",
+                        "title": story_title,
+                        "description": story_description,
+                        "t_shirt_size": story_tshirt,
+                        "story_points": story_points,
+                        "wsjf_score": story_wsjf,
+                        "ato_impact_tier": story_ato,
+                        "parent_id": feature_id,
+                        "status": "draft",
+                    }
+                    if generate_bdd:
+                        story_item["acceptance_criteria"] = story_bdd
+                    items.append(story_item)
+                    counts["stories"] += 1
+
+                    # Update requirement status to decomposed
+                    conn.execute(
+                        "UPDATE intake_requirements SET status = 'decomposed', updated_at = ? WHERE id = ?",
+                        (datetime.now().isoformat(), req["id"]),
+                    )
+
+            # --- Create Enabler for infrastructure needs of this epic ---
+            if target_level in ("story", "enabler"):
+                enabler_id = _generate_id("enbl")
+                enabler_title = f"Infrastructure Enabler: {rtype.replace('_', ' ').title()}"
+                enabler_description = (
+                    f"Infrastructure and platform enablement for {rtype} capabilities. "
+                    f"Covers environment setup, CI/CD pipeline configuration, "
+                    f"security hardening, and compliance scaffolding required to "
+                    f"support {len(group_reqs)} {rtype} requirement(s)."
+                )
+                enabler_tshirt = _estimate_tshirt(len(group_reqs)) if estimate else None
+                enabler_wsjf = (
+                    _compute_wsjf("high", enabler_tshirt)
+                    if estimate and enabler_tshirt else None
+                )
+
+                enabler_bdd = None
+                if generate_bdd:
+                    enabler_bdd = generate_bdd_criteria(enabler_description, rtype)
+
+                conn.execute(
+                    """INSERT INTO safe_decomposition
+                       (id, session_id, project_id, parent_id, level, title,
+                        description, acceptance_criteria, t_shirt_size, story_points,
+                        pi_target, wsjf_score, ato_impact_tier, source_requirement_ids,
+                        status, classification, created_at)
+                       VALUES (?, ?, ?, ?, 'enabler', ?, ?, ?, ?, NULL, NULL, ?, ?, ?, 'draft', 'CUI', ?)""",
+                    (
+                        enabler_id, session_id, project_id, epic_id,
+                        enabler_title, enabler_description, enabler_bdd,
+                        enabler_tshirt, enabler_wsjf, epic_ato,
+                        json.dumps(req_ids),
+                        datetime.now().isoformat(),
+                    ),
+                )
+
+                enabler_item = {
+                    "id": enabler_id,
+                    "level": "enabler",
+                    "title": enabler_title,
+                    "description": enabler_description,
+                    "t_shirt_size": enabler_tshirt,
+                    "wsjf_score": enabler_wsjf,
+                    "ato_impact_tier": epic_ato,
+                    "parent_id": epic_id,
+                    "status": "draft",
+                }
+                if generate_bdd:
+                    enabler_item["acceptance_criteria"] = enabler_bdd
+                items.append(enabler_item)
+                counts["enablers"] += 1
+
+    # Update session decomposed count
+    total_items = sum(counts.values())
+    conn.execute(
+        "UPDATE intake_sessions SET decomposed_count = ?, updated_at = ? WHERE id = ?",
+        (total_items, datetime.now().isoformat(), session_id),
+    )
+
+    conn.commit()
+    conn.close()
+
+    if _HAS_AUDIT:
+        log_event(
+            event_type="decomposition_generated",
+            actor="icdev-requirements-analyst",
+            action=f"Decomposed {len(reqs)} requirements into {total_items} SAFe items",
+            project_id=project_id,
+            details={
+                "session_id": session_id,
+                "items_created": total_items,
+                "levels": counts,
+            },
+        )
+
+    return {
+        "status": "ok",
+        "session_id": session_id,
+        "items_created": total_items,
+        "levels": counts,
+        "items": items,
+    }
+
+
+# ---------------------------------------------------------------------------
+# Query existing decomposition
+# ---------------------------------------------------------------------------
+
+def get_decomposition(session_id, level=None, db_path=None):
+    """Return existing decomposition items for a session.
+
+    Args:
+        session_id: Intake session ID.
+        level: Optional filter by SAFe level (epic, feature, story, enabler).
+        db_path: Optional DB path override.
+
+    Returns:
+        dict with session_id and list of decomposition items.
+    """
+    conn = _get_connection(db_path)
+
+    session = conn.execute(
+        "SELECT * FROM intake_sessions WHERE id = ?", (session_id,)
+    ).fetchone()
+    if not session:
+        conn.close()
+        raise ValueError(f"Session '{session_id}' not found.")
+
+    if level:
+        rows = conn.execute(
+            "SELECT * FROM safe_decomposition WHERE session_id = ? AND level = ? ORDER BY created_at",
+            (session_id, level),
+        ).fetchall()
+    else:
+        rows = conn.execute(
+            "SELECT * FROM safe_decomposition WHERE session_id = ? ORDER BY created_at",
+            (session_id,),
+        ).fetchall()
+
+    items = [dict(r) for r in rows]
+    conn.close()
+
+    # Compute level counts
+    level_counts = {"epics": 0, "features": 0, "stories": 0, "enablers": 0}
+    for item in items:
+        lv = item.get("level", "")
+        if lv == "epic":
+            level_counts["epics"] += 1
+        elif lv == "feature":
+            level_counts["features"] += 1
+        elif lv == "story":
+            level_counts["stories"] += 1
+        elif lv == "enabler":
+            level_counts["enablers"] += 1
+
+    return {
+        "status": "ok",
+        "session_id": session_id,
+        "filter_level": level,
+        "total_items": len(items),
+        "levels": level_counts,
+        "items": items,
+    }
+
+
+# ---------------------------------------------------------------------------
+# Parallel group detection (D161, spec-kit Pattern 7)
+# ---------------------------------------------------------------------------
+
+_DEPENDENCY_KEYWORDS = [
+    "after", "depends on", "requires", "once", "from step",
+    "building on", "following", "result of", "output of",
+    "blocked by", "prerequisite", "prior to",
+]
+
+
+def detect_parallel_groups(session_id, db_path=None):
+    """Detect groups of independent SAFe items that can execute in parallel.
+
+    Items at the same level under the same parent with no cross-dependencies
+    are grouped together. Returns list of group dicts.
+
+    ADR D161: Reuses existing DAG infrastructure for concurrency annotation.
+    """
+    conn = _get_connection(db_path)
+    items = conn.execute(
+        """SELECT id, title, level, parent_id, acceptance_criteria,
+                  parallel_group
+           FROM safe_decomposition
+           WHERE session_id = ?
+           ORDER BY level, parent_id, created_at""",
+        (session_id,),
+    ).fetchall()
+    items = [dict(r) for r in items]
+    conn_write = conn  # reuse for updates
+
+    if not items:
+        conn.close()
+        return []
+
+    # Group items by (parent_id, level)
+    from collections import defaultdict
+    siblings = defaultdict(list)
+    for item in items:
+        key = (item.get("parent_id") or "_root", item.get("level", ""))
+        siblings[key].append(item)
+
+    groups = []
+    group_index = 0
+
+    for (parent_id, level), sibling_list in siblings.items():
+        if len(sibling_list) < 2:
+            continue  # Need at least 2 siblings for parallelism
+
+        # Check each item for dependency keywords in title + acceptance_criteria
+        independent = []
+        dependent = []
+        for item in sibling_list:
+            text = (item.get("title", "") + " " +
+                    (item.get("acceptance_criteria") or "")).lower()
+            has_dep = any(kw in text for kw in _DEPENDENCY_KEYWORDS)
+            if has_dep:
+                dependent.append(item)
+            else:
+                independent.append(item)
+
+        if len(independent) >= 2:
+            group_id = f"pg-{(parent_id or 'root')[:8]}-{group_index}"
+            group_index += 1
+            for item in independent:
+                conn_write.execute(
+                    "UPDATE safe_decomposition SET parallel_group = ? WHERE id = ?",
+                    (group_id, item["id"]),
+                )
+            groups.append({
+                "group_id": group_id,
+                "parent_id": parent_id,
+                "level": level,
+                "item_count": len(independent),
+                "items": [{"id": i["id"], "title": i["title"]} for i in independent],
+            })
+
+    conn_write.commit()
+    conn.close()
+    return groups
+
+
+# ---------------------------------------------------------------------------
+# CLI
+# ---------------------------------------------------------------------------
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="ICDEV SAFe Decomposition Engine"
+    )
+    parser.add_argument("--session-id", required=True, help="Intake session ID")
+    parser.add_argument(
+        "--level",
+        choices=["epic", "capability", "feature", "story", "enabler"],
+        default="story",
+        help="Target decomposition level (default: story)",
+    )
+    parser.add_argument(
+        "--generate-bdd", action="store_true",
+        help="Generate BDD Given/When/Then acceptance criteria",
+    )
+    parser.add_argument(
+        "--estimate", action="store_true", default=True,
+        help="Add T-shirt size estimates and WSJF scores (default: True)",
+    )
+    parser.add_argument(
+        "--get", action="store_true",
+        help="Get existing decomposition instead of generating",
+    )
+    parser.add_argument(
+        "--annotate-parallel", action="store_true",
+        help="Detect and annotate parallel task groups (D161)",
+    )
+    parser.add_argument("--json", action="store_true", help="JSON output")
+    args = parser.parse_args()
+
+    try:
+        if args.annotate_parallel:
+            groups = detect_parallel_groups(args.session_id)
+            result = {
+                "status": "ok",
+                "session_id": args.session_id,
+                "parallel_groups": groups,
+                "total_groups": len(groups),
+            }
+            if args.json:
+                print(json.dumps(result, indent=2, default=str))
+            else:
+                print(f"Detected {len(groups)} parallel group(s)")
+                for g in groups:
+                    print(f"  [{g['group_id']}] {g['item_count']} items at {g['level']} level")
+                    for item in g["items"]:
+                        print(f"    - {item['title']}")
+            return
+
+        if args.get:
+            # Query existing decomposition
+            # If --level is explicitly provided with --get, use it; otherwise None
+            # We detect explicit --level by checking if it differs from default
+            # For --get without --level, return all levels
+            import sys
+            explicit_level = None
+            if "--level" in sys.argv:
+                explicit_level = args.level
+            result = get_decomposition(
+                args.session_id, level=explicit_level,
+            )
+        else:
+            # Generate decomposition
+            result = decompose_requirements(
+                args.session_id,
+                target_level=args.level,
+                generate_bdd=args.generate_bdd,
+                estimate=args.estimate,
+            )
+
+        if args.json:
+            print(json.dumps(result, indent=2, default=str))
+        else:
+            if args.get:
+                print(f"Decomposition for session {args.session_id}: "
+                      f"{result['total_items']} items")
+                for item in result.get("items", []):
+                    level = item.get("level", "?")
+                    indent = "  " * (
+                        ["epic", "capability", "feature", "story", "enabler"].index(level)
+                        if level in ["epic", "capability", "feature", "story", "enabler"]
+                        else 0
+                    )
+                    print(f"{indent}[{level.upper()}] {item.get('title', '?')}")
+            else:
+                print(
+                    f"Created {result['items_created']} SAFe items "
+                    f"from session {args.session_id}"
+                )
+                levels = result.get("levels", {})
+                print(
+                    f"  Epics: {levels.get('epics', 0)}, "
+                    f"Features: {levels.get('features', 0)}, "
+                    f"Stories: {levels.get('stories', 0)}, "
+                    f"Enablers: {levels.get('enablers', 0)}"
+                )
+                for item in result.get("items", []):
+                    level = item.get("level", "?")
+                    indent = "  " * (
+                        ["epic", "capability", "feature", "story", "enabler"].index(level)
+                        if level in ["epic", "capability", "feature", "story", "enabler"]
+                        else 0
+                    )
+                    print(f"{indent}[{level.upper()}] {item.get('title', '?')}")
+
+    except (ValueError, FileNotFoundError) as e:
+        if args.json:
+            print(json.dumps({"error": str(e)}, indent=2))
+        else:
+            print(f"Error: {e}")
+        raise SystemExit(1)
+
+
+if __name__ == "__main__":
+    main()
+# [TEMPLATE: CUI // SP-CTI]