Team Ultra Analyze

Team Ultra Analyze

Deep collaborative analysis: explore -> analyze -> discuss -> synthesize. Supports Quick/Standard/Deep pipeline modes with configurable depth (N parallel agents). Discussion loops enable user-guided progressive understanding.

Architecture

Skill(skill="team-ultra-analyze", args="<topic>")
                    |
         SKILL.md (this file) = Router
                    |
     +--------------+--------------+
     |                             |
  no --role flag              --role <name>
     |                             |
  Coordinator                  Worker
  roles/coordinator/role.md    roles/<name>/role.md
     |
     +-- analyze -> dispatch -> spawn workers -> STOP
                                    |
                    +-------+-------+-------+-------+
                    v       v       v       v
                 [team-worker agents, each loads roles/<role>/role.md]

Pipeline (Standard mode):
  [EXPLORE-1..N](parallel) -> [ANALYZE-1..N](parallel) -> DISCUSS-001 -> SYNTH-001

Pipeline (Deep mode):
  [EXPLORE-1..N] -> [ANALYZE-1..N] -> DISCUSS-001 -> ANALYZE-fix -> DISCUSS-002 -> ... -> SYNTH-001

Pipeline (Quick mode):
  EXPLORE-001 -> ANALYZE-001 -> SYNTH-001

Role Registry

Role Router

Parse $ARGUMENTS:

• Has --role <name> → Read roles/<name>/role.md, execute Phase 2-4
• No --role → roles/coordinator/role.md, execute entry router

Delegation Lock

Coordinator is a PURE ORCHESTRATOR. It coordinates, it does NOT do.

Before calling ANY tool, apply this check:

If a tool call is BLOCKED: STOP. Create a task, spawn a worker.

No exceptions for "simple" tasks. Even a single-file read-and-report MUST go through spawn_agent.

Shared Constants

• Session prefix: UAN
• Session path: .workflow/.team/UAN-<date>-<slug>/
• Team name: ultra-analyze
• CLI tools: ccw cli --mode analysis (read-only), ccw cli --mode write (modifications)
• Message bus: mcp__ccw-tools__team_msg(session_id=<session-id>, ...)

Worker Spawn Template

Coordinator spawns workers using this template:

spawn_agent({
  agent_type: "team_worker",
  task_name: "<task-id>",
  fork_turns: "none",
  message: `## Role Assignment
role: <role>
role_spec: <skill_root>/roles/<role>/role.md
session: <session-folder>
session_id: <session-id>
requirement: <topic-description>
inner_loop: false

Read role_spec file (<skill_root>/roles/<role>/role.md) to load Phase 2-4 domain instructions.

## Task Context
task_id: <task-id>
title: <task-title>
description: <task-description>
pipeline_phase: <pipeline-phase>

## Upstream Context
<prev_context>`
})

After spawning, use wait_agent({ timeout_ms: 1800000 }) to collect results. If result.timed_out, send STATUS_CHECK via followup_task (wait 3 min), then FINALIZE with interrupt (wait 3 min), then mark timed_out and close agents. Use close_agent({ target }) each worker.

Model Selection Guide

Override model/reasoning_effort in spawn_agent when cost optimization is needed:

spawn_agent({
  agent_type: "team_worker",
  task_name: "<task-id>",
  fork_turns: "none",
  model: "<model-override>",
  reasoning_effort: "<effort-level>",
  message: "..."
})

User Commands

Session Directory

.workflow/.team/UAN-{YYYY-MM-DD}-{slug}/
+-- .msg/messages.jsonl          # Message bus log
+-- .msg/meta.json               # Session metadata + cross-role state
+-- discussion.md                # Understanding evolution and discussion timeline
+-- explorations/                # Explorer output
|   +-- exploration-001.json
|   +-- exploration-002.json
+-- analyses/                    # Analyst output
|   +-- analysis-001.json
|   +-- analysis-002.json
+-- discussions/                 # Discussant output
|   +-- discussion-round-001.json
+-- conclusions.json             # Synthesizer output
+-- wisdom/                      # Cross-task knowledge
|   +-- learnings.md
|   +-- decisions.md
|   +-- conventions.md
|   +-- issues.md

v4 Agent Coordination

Message Semantics

Parallel Phase Coordination

Standard/Deep modes spawn N parallel agents in EXPLORE and ANALYZE phases. Use batch spawn + wait:

// EXPLORE phase: spawn N explorers in parallel
const explorerNames = ["EXPLORE-001", "EXPLORE-002", ..., "EXPLORE-00N"]
for (const name of explorerNames) {
  spawn_agent({ agent_type: "team_worker", task_name: name, ... })
}
wait_agent({ timeout_ms: 1800000 })  // 30 min — apply timeout cascade if timed_out
// Collect all results, then spawn ANALYZE phase

// ANALYZE phase: send exploration context to analysts via items (not send_message)
// since analysts are spawned AFTER explorers complete

Agent Health Check

Use list_agents({}) in handleResume and handleComplete:

// Reconcile session state with actual running agents
const running = list_agents({})
// Compare with tasks.json active_agents
// Reset orphaned tasks (in_progress but agent gone) to pending
// Critical for parallel phases -- multiple agents may crash independently

Named Agent Targeting

Workers are spawned with task_name: "<task-id>" enabling direct addressing:

• send_message({ target: "ANALYZE-001", message: "..." }) -- queue supplementary exploration findings
• close_agent({ target: "EXPLORE-001" }) -- cleanup by name after wait_agent returns

Completion Action

When pipeline completes, coordinator presents:

functions.request_user_input({
  questions: [{
    question: "Ultra-Analyze pipeline complete. What would you like to do?",
    header: "Completion",
    multiSelect: false,
    options: [
      { label: "Archive & Clean (Recommended)", description: "Archive session, clean up tasks and team resources" },
      { label: "Keep Active", description: "Keep session active for follow-up work or inspection" },
      { label: "Export Results", description: "Export deliverables to a specified location, then clean" }
    ]
  }]
})

Specs Reference

• specs/team-config.json — Team configuration and pipeline settings

Error Handling