AgentSkillsCN

workflow

以能力为导向的核心“邪恶团队”v3工作流。智能决策模块会根据输入信息,精准识别并调配所需的专业人才;若遇专业人才暂时缺位,系统将自动启用内置的兜底方案。 当用户提出“启动一个项目”、“明确项目成果”、“进入设计阶段”、“质量工程关卡”、“左移测试”、“直接完成”、“进入审批阶段”,或亟需一套结构化的软件交付工作流指导时,可选用此方案。

SKILL.md
--- frontmatter
name: workflow
description: |
  Core wicked-crew v3 workflow with capability-based orchestration.
  Smart decisioning analyzes input to determine specialists needed.
  Built-in fallbacks when specialists unavailable.

  Use when the user mentions "start a project", "clarify outcome", "design phase",
  "QE gate", "shift-left testing", "just finish", "approve phase", or needs structured
  software delivery workflow guidance.

Workflow Skill (v3)

Capability-based orchestration with smart specialist engagement.

v3 Architecture

code
User Input → Smart Decisioning → Specialist Discovery → Engagement
                  │                     │
                  ├── Signal Detection  ├── Available specialists
                  ├── Complexity Score  ├── Personas
                  └── Ambiguity Check   └── Fallback agents

Phase Progression

code
clarify → design → qe → build → review

Each phase has: Clear deliverables, specialist engagement based on signals, approval gate.

Smart Decisioning

Signal Categories

SignalKeywordsSpecialists
Securityauth, encrypt, token, jwtdevsecops, compliance
Performancescale, optimize, cacheappeng, qe
Productuser, feature, storyproduct, strategy
ComplianceSOC2, HIPAA, auditcompliance
Ambiguitymaybe, should we, optionsjam
Complexityintegration, migrate, refactorpmo, arch

Complexity Scoring (0-7)

  • 0-2: Simple → Built-in agents only
  • 3-4: Moderate → Core specialists
  • 5-7: Complex → All relevant + PMO

Specialist Engagement

Discovery

Crew discovers specialists via specialist.json in each plugin.

Fallback Agents

SpecialistFallback
jamfacilitator
qe, strategyreviewer
archresearcher
appeng, devsecopsimplementer

Phase Details

Clarify

Goal: Define success criteria Deliverables: Outcome statement, success criteria, scope boundaries Specialists: jam (if ambiguous), product

Design

Goal: Architect the solution Deliverables: Architecture docs, pattern identification, technical approach Specialists: strategy, arch, appeng

QE (Quality Engineering)

Goal: Define test strategy before building Deliverables: Test scenarios, risk assessment, edge cases Specialists: qe, devsecops (if security signals)

Build

Goal: Implement the solution Deliverables: Working implementation, progress tracking, tests passing Specialists: appeng, devsecops (if infra)

Review

Goal: Multi-perspective validation Deliverables: Review findings, recommendations, sign-off Specialists: strategy, qe, compliance (if regulated)

Commands Reference

CommandPurpose
/wicked-crew:startBegin project with smart decisioning
/wicked-crew:statusView current state and engaged specialists
/wicked-crew:executeRun current phase with specialists
/wicked-crew:approveApprove and advance phase
/wicked-crew:just-finishAutonomous completion
/wicked-crew:gateRun QE gate (value/strategy/execution)
/wicked-crew:evidenceQuery evidence for a task

Event Flow

Project Start

code
crew:project:started:success
crew:specialist:engaged:success
crew:specialist:unavailable:warning (with fallback)

Phase Transitions

code
crew:phase:started:success
crew:phase:completed:success
crew:phase:approved:success

Utility Integration

PluginEnhancementFallback
wicked-kanbanPersistent task boardTodoWrite
wicked-memCross-session learningProject files
wicked-cacheCached analysisRe-run each time

See Integration Details for usage patterns.

Evidence Tracking

Gate decisions and artifacts are tracked as evidence. See EVIDENCE.md for:

  • Evidence tiers (L1-L4)
  • Artifact naming conventions
  • Automatic vs manual evidence collection

Configuration

Customize in ~/.something-wicked/wicked-crew/config.yaml:

yaml
defaults:
  always_engage: [qe]      # Always use QE specialist
  complexity_threshold: 4  # Lower = more specialists