Jira Issue Triage and Routing Skill

Intelligent classification, prioritization, and routing system for Jira issues.

Triage Decision Tree

START: New Jira Issue
├─ Step 1: ISSUE TYPE CLASSIFICATION
│  ├─ Bug? → Classify severity (Blocker/Critical/Major/Minor) → HIGH Priority if Critical
│  ├─ Story/Feature? → Check requirements clarity, sprint scope, dependencies
│  ├─ Epic? → Route to epic-decomposer (NEVER implement directly)
│  ├─ Task? → Route by category (Tech Debt/Config/Docs/Infrastructure)
│  └─ Spike? → Time-box (1-2 days max) → Document findings → Create stories
│
├─ Step 2: COMPLEXITY ASSESSMENT (0-100 scale)
│  ├─ Code Impact: Single file (2) → Multiple services (10)
│  ├─ Integration: None (0) → Multiple external APIs (10)
│  ├─ Risk: None (0) → Critical/Data loss (10)
│  ├─ Testing: No tests (0) → Complex E2E scenarios (10)
│  ├─ Dependencies: None (0) → Multiple blocking (10)
│  └─ Uncertainty: Known (0) → Complete unknown (10)
│
│  Scoring: (Code×0.25 + Integration×0.20 + Risk×0.20 + Testing×0.15 + Dependencies×0.10 + Uncertainty×0.10) × 10
│
│  Categories:
│  • 1-20: SIMPLE → Quick-Fix Path (2-3 agents, 2-4 hrs)
│  • 21-40: MODERATE → Standard Workflow (3-5 agents, 2-5 days)
│  • 41-70: COMPLEX → Extended Workflow (5-10 agents, 5-10 days)
│  • 71+: VERY COMPLEX → Decomposition Path
│
├─ Step 3: PRIORITY & SEVERITY ASSESSMENT
│  ├─ Business Impact: Blocks production (BLOCKER) → Nice-to-have (LOW)
│  ├─ Urgency: Immediate (hours) → Backlog (months)
│  └─ Bug Severity Matrix:
│     • BLOCKER: Production down, data loss, security breach (1-4 hr SLA)
│     • CRITICAL: Major functionality broken (4-8 hr SLA)
│     • MAJOR: Important feature degraded (1-3 days)
│     • MINOR: Cosmetic issues (next sprint)
│
├─ Step 4: WORKFLOW ROUTING
│  ├─ QUICK-FIX: Simple (1-20), <50 LOC, low risk → EXPLORE→CODE→TEST→COMMIT
│  ├─ STANDARD: Moderate (21-40) → EXPLORE→PLAN→CODE→TEST→FIX→COMMIT
│  ├─ EXTENDED: Complex (41-70), high risk → All phases + extended thinking + checkpoints
│  ├─ RESEARCH: Spike/POC → RESEARCH→DOCUMENT→CREATE STORIES (time-boxed)
│  └─ DECOMPOSITION: Epic (71+) → ANALYZE→DECOMPOSE→CREATE STORIES→TRIAGE EACH
│
├─ Step 5: AGENT SELECTION
│  ├─ By Type: Bug→debugger; Story→feature-architect; Epic→epic-decomposer; Spike→research-specialist
│  ├─ By Tech: Frontend→react-specialist; Backend→nodejs/python/java-specialist; DevOps→k8s-specialist
│  ├─ By Phase: EXPLORE→code-analyst; PLAN→architect; CODE→tech-specific; TEST→test-engineer; COMMIT→git-specialist
│  └─ By Complexity: SIMPLE→2-3 junior agents; COMPLEX→5-10 seniors + extended thinking
│
├─ Step 6: RISK & ESCALATION
│  ├─ Risk Factors: Security, breaking changes, data migrations, compliance issues
│  ├─ Level 1 (IMMEDIATE STOP): Security vulnerability, data loss, compliance breach
│  ├─ Level 2 (CHECKPOINT): Complexity +50%, blocker >4hrs, breaking changes
│  └─ Level 3 (POST-COMPLETION): Standard bugs, docs, minor refactoring
│
└─ Step 7: OUTPUT ROUTING PACKAGE
   ├─ Classification, complexity score, priority, workflow path
   ├─ Agent selection, risk assessment, escalation triggers
   └─ Update Jira labels, assign workflow, spawn agents

Issue Type Routing Matrix

Complexity Scoring Example

Issue: Add CSV export functionality

Code Impact: 5 (3-4 files, 200-300 LOC) × 0.25 = 1.25
Integration: 3 (internal APIs) × 0.20 = 0.60
Risk: 4 (file generation, perf) × 0.20 = 0.80
Testing: 6 (unit + integration) × 0.15 = 0.90
Dependencies: 2 (DB query opt) × 0.10 = 0.20
Uncertainty: 3 (format details) × 0.10 = 0.30

Score: 3.65 × 10 = 36.5 → MODERATE (5 story points)
→ Standard Workflow, 3-5 agents, 2-5 days

Epic Decomposition Example

Epic: User Management System

Phase 1 (Sprint 1): PROJ-101 Profile CRUD (5pts), PROJ-102 RBAC (8pts), PROJ-103 Password Policy (3pts)
Phase 2 (Sprint 2): PROJ-104 Audit Logging (5pts), PROJ-105 MFA (8pts)
Phase 3 (Sprint 3): PROJ-106 Bulk Ops (5pts), PROJ-107 Import/Export (5pts)
Phase 4 (Sprint 4): PROJ-108 Analytics (8pts)

Dependencies: PROJ-102 blocks PROJ-106; PROJ-101 blocks PROJ-104
Total: 8 stories, 47 points, 4 sprints, 2-3 developers

Workflow Paths

Agent Selection by Phase

Escalation Matrix

Bug Routing Decision

Bug Detected
├─ Can reproduce consistently? NO → Gather reproduction steps
├─ Security vulnerability? YES → IMMEDIATE ESCALATION
├─ Severity?
│  ├─ BLOCKER/CRITICAL → Extended workflow, root-cause analysis, human checkpoints
│  ├─ MAJOR → Standard workflow, add regression tests
│  └─ MINOR → Quick-Fix path, batch with similar fixes
└─ Root cause known? NO → Extend EXPLORE phase with debugger/profiler

Feature Routing Decision

Story/Feature Detected
├─ Requirements clear? NO → Route to requirements-analyzer first
├─ Single sprint scope? NO → Consider epic decomposition
├─ Tech complexity?
│  ├─ Frontend → UI specialists
│  ├─ Backend → API specialists
│  ├─ Full-stack → Both
│  └─ Infrastructure → DevOps specialists
├─ Integration complexity?
│  ├─ None → SIMPLE (1-20 pts)
│  ├─ Internal APIs → MODERATE (21-40 pts)
│  ├─ External APIs → COMPLEX (41-70 pts)
│  └─ Multiple external → VERY COMPLEX (71+ pts)
└─ Score 1-100 and route to appropriate workflow

Spike Handling

Spike Detected
├─ Time-Box Definition (CRITICAL)
│  ├─ Small: 1 day max
│  ├─ Medium: 2-3 days max
│  └─ Large: 1 week max (justify)
├─ Research Phase: Investigate tech/approach, build POC if needed
├─ Document Phase: Write findings, recommend approach, estimate effort
└─ Output: Go/No-Go decision OR implementation stories with estimates

Triage Output Checklist

• Issue classification (type, subtype, confidence)
• Complexity score (0-100) with factor breakdown
• Priority & severity assessment
• Workflow path selection (Quick-Fix/Standard/Extended/Research/Decomposition)
• Agent recommendations per phase
• Risk assessment and escalation level
• Dependencies and blockers
• Jira labels and field updates
• Execute workflow with selected agents