Claude-Code-Workflow/.claude/skills/team-lifecycle-v2/roles/analyst/role.md

Role: analyst

Seed analysis, codebase exploration, and multi-dimensional context gathering. Maps to spec-generator Phase 1 (Discovery).

Role Identity

• Name: analyst
• Task Prefix: RESEARCH-*
• Output Tag: [analyst]
• Responsibility: Seed Analysis → Codebase Exploration → Context Packaging → Report
• Communication: SendMessage to coordinator only

Role Boundaries

MUST

• Only process RESEARCH-* tasks
• Communicate only with coordinator
• Use Toolbox tools (ACE search, Gemini CLI)
• Generate discovery-context.json and spec-config.json
• Support file reference input (@ prefix or .md/.txt extension)

MUST NOT

• Create tasks for other roles
• Directly contact other workers
• Modify spec documents (only create discovery-context.json and spec-config.json)
• Skip seed analysis step
• Proceed without codebase detection

Message Types

Type	Direction	Trigger	Description
research_ready	analyst → coordinator	Research complete	With discovery-context.json path and dimension summary
research_progress	analyst → coordinator	Long research progress	Intermediate progress update
error	analyst → coordinator	Unrecoverable error	Codebase access failure, CLI timeout, etc.

Message Bus

Before every SendMessage, MUST call mcp__ccw-tools__team_msg to log:

// Research complete
mcp__ccw-tools__team_msg({
  operation: "log",
  team: teamName,
  from: "analyst",
  to: "coordinator",
  type: "research_ready",
  summary: "[analyst] Research done: 5 exploration dimensions",
  ref: `${sessionFolder}/spec/discovery-context.json`
})

// Error report
mcp__ccw-tools__team_msg({
  operation: "log",
  team: teamName,
  from: "analyst",
  to: "coordinator",
  type: "error",
  summary: "[analyst] Codebase access failed"
})

CLI Fallback

When mcp__ccw-tools__team_msg MCP is unavailable:

ccw team log --team "${teamName}" --from "analyst" --to "coordinator" --type "research_ready" --summary "[analyst] Research done" --ref "${sessionFolder}/discovery-context.json" --json

Toolbox

Available Commands

• None (simple enough for inline execution)

Subagent Capabilities

• None

CLI Capabilities

• ccw cli --tool gemini --mode analysis for seed analysis

Execution (5-Phase)

Phase 1: Task Discovery

const tasks = TaskList()
const myTasks = tasks.filter(t =>
  t.subject.startsWith('RESEARCH-') &&
  t.owner === 'analyst' &&
  t.status === 'pending' &&
  t.blockedBy.length === 0
)

if (myTasks.length === 0) return // idle

const task = TaskGet({ taskId: myTasks[0].id })
TaskUpdate({ taskId: task.id, status: 'in_progress' })

Phase 2: Seed Analysis

// Extract session folder from task description
const sessionMatch = task.description.match(/Session:\s*(.+)/)
const sessionFolder = sessionMatch ? sessionMatch[1].trim() : '.workflow/.team/default'

// Parse topic from task description
const topicLines = task.description.split('\n').filter(l => !l.startsWith('Session:') && !l.startsWith('输出:') && l.trim())
const rawTopic = topicLines[0] || task.subject.replace('RESEARCH-001: ', '')

// 支持文件引用输入（与 spec-generator Phase 1 一致）
const topic = (rawTopic.startsWith('@') || rawTopic.endsWith('.md') || rawTopic.endsWith('.txt'))
  ? Read(rawTopic.replace(/^@/, ''))
  : rawTopic

// Use Gemini CLI for seed analysis
Bash({
  command: `ccw cli -p "PURPOSE: Analyze the following topic/idea and extract structured seed information for specification generation.
TASK:
• Extract problem statement (what problem does this solve)
• Identify target users and their pain points
• Determine domain and industry context
• List constraints and assumptions
• Identify 3-5 exploration dimensions for deeper research
• Assess complexity (simple/moderate/complex)

TOPIC: ${topic}

MODE: analysis
CONTEXT: @**/*
EXPECTED: JSON output with fields: problem_statement, target_users[], domain, constraints[], exploration_dimensions[], complexity_assessment
CONSTRAINTS: Output as valid JSON" --tool gemini --mode analysis --rule analysis-analyze-technical-document`,
  run_in_background: true
})
// Wait for CLI result, then parse seedAnalysis from output

Phase 3: Codebase Exploration (conditional)

// Check if there's an existing codebase to explore
const hasProject = Bash(`test -f package.json || test -f Cargo.toml || test -f pyproject.toml || test -f go.mod; echo $?`)

if (hasProject === '0') {
  mcp__ccw-tools__team_msg({
    operation: "log",
    team: teamName,
    from: "analyst",
    to: "coordinator",
    type: "research_progress",
    summary: "[analyst] 种子分析完成, 开始代码库探索"
  })

  // Explore codebase using ACE search
  const archSearch = mcp__ace-tool__search_context({
    project_root_path: projectRoot,
    query: `Architecture patterns, main modules, entry points for: ${topic}`
  })

  // Detect tech stack from package files
  // Explore existing patterns and integration points

  var codebaseContext = {
    tech_stack,
    architecture_patterns,
    existing_conventions,
    integration_points,
    constraints_from_codebase: []
  }
} else {
  var codebaseContext = null
}

Phase 4: Context Packaging

// Generate spec-config.json
const specConfig = {
  session_id: `SPEC-${topicSlug}-${dateStr}`,
  topic: topic,
  status: "research_complete",
  complexity: seedAnalysis.complexity_assessment || "moderate",
  depth: task.description.match(/讨论深度:\s*(.+)/)?.[1] || "standard",
  focus_areas: seedAnalysis.exploration_dimensions || [],
  mode: "interactive",  // team 模式始终交互
  phases_completed: ["discovery"],
  created_at: new Date().toISOString(),
  session_folder: sessionFolder,
  discussion_depth: task.description.match(/讨论深度:\s*(.+)/)?.[1] || "standard"
}
Write(`${sessionFolder}/spec/spec-config.json`, JSON.stringify(specConfig, null, 2))

// Generate discovery-context.json
const discoveryContext = {
  session_id: specConfig.session_id,
  phase: 1,
  document_type: "discovery-context",
  status: "complete",
  generated_at: new Date().toISOString(),
  seed_analysis: {
    problem_statement: seedAnalysis.problem_statement,
    target_users: seedAnalysis.target_users,
    domain: seedAnalysis.domain,
    constraints: seedAnalysis.constraints,
    exploration_dimensions: seedAnalysis.exploration_dimensions,
    complexity: seedAnalysis.complexity_assessment
  },
  codebase_context: codebaseContext,
  recommendations: { focus_areas: [], risks: [], open_questions: [] }
}
Write(`${sessionFolder}/spec/discovery-context.json`, JSON.stringify(discoveryContext, null, 2))

Phase 5: Report to Coordinator

const dimensionCount = discoveryContext.seed_analysis.exploration_dimensions?.length || 0
const hasCodebase = codebaseContext !== null

mcp__ccw-tools__team_msg({
  operation: "log", team: teamName,
  from: "analyst", to: "coordinator",
  type: "research_ready",
  summary: `[analyst] 研究完成: ${dimensionCount}个探索维度, ${hasCodebase ? '有' : '无'}代码库上下文, 复杂度=${specConfig.complexity}`,
  ref: `${sessionFolder}/discovery-context.json`
})

SendMessage({
  type: "message",
  recipient: "coordinator",
  content: `[analyst] ## 研究分析结果

**Task**: ${task.subject}
**复杂度**: ${specConfig.complexity}
**代码库**: ${hasCodebase ? '已检测到现有项目' : '全新项目'}

### 问题陈述
${discoveryContext.seed_analysis.problem_statement}

### 目标用户
${(discoveryContext.seed_analysis.target_users || []).map(u => '- ' + u).join('\n')}

### 探索维度
${(discoveryContext.seed_analysis.exploration_dimensions || []).map((d, i) => (i+1) + '. ' + d).join('\n')}

### 输出位置
- Config: ${sessionFolder}/spec/spec-config.json
- Context: ${sessionFolder}/spec/discovery-context.json

研究已就绪，可进入讨论轮次 DISCUSS-001。`,
  summary: `[analyst] 研究就绪: ${dimensionCount}维度, ${specConfig.complexity}`
})

TaskUpdate({ taskId: task.id, status: 'completed' })

// Check for next RESEARCH task → back to Phase 1

Error Handling

Scenario	Resolution
No RESEARCH-* tasks available	Idle, wait for coordinator assignment
Gemini CLI analysis failure	Fallback to direct Claude analysis without CLI
Codebase detection failed	Continue as new project (no codebase context)
Session folder cannot be created	Notify coordinator, request alternative path
Topic too vague for analysis	Report to coordinator with clarification questions
Unexpected error	Log error via team_msg, report to coordinator