# Coordinator Mode

> Multi-agent orchestration patterns: when to use coordinator mode, worker prompts, parallel execution

- Skill: `sefaertunc/coordinator-mode` (Agent Skill)
- Install (CLI): `npx skillmds@latest add sefaertunc/coordinator-mode`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sefaertunc/coordinator-mode/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: sefaertunc (https://skillmd.com/u/sefaertunc)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/sefaertunc/coordinator-mode

---


# Coordinator Mode

## When to Use Multi-Agent Coordination

Use coordinator patterns when:

- A task has independent research + implementation phases
- Multiple file areas can be worked on in parallel
- Verification should run alongside implementation
- You need to synthesize findings before directing follow-up work

Don't use coordinator patterns when:

- The task is small enough for one session
- Steps are strictly sequential with no parallelism
- The overhead of coordination exceeds the work itself

## Phase-Based Workflow

Most coordinated tasks follow four phases:

| Phase | Who | Purpose |
|-------|-----|---------|
| Research | Workers (parallel) | Investigate codebase, find files, understand problem |
| Synthesis | **You** (coordinator) | Read findings, understand the problem, craft implementation specs |
| Implementation | Workers | Make targeted changes per spec, commit |
| Verification | Workers | Test changes work, try to break them |

The synthesis phase is the coordinator's most important job. When workers report
research findings, **you must understand them before directing follow-up work**.
Read the findings. Identify the approach. Then write a prompt that proves you
understood by including specific file paths, line numbers, and exactly what to change.

Never write "based on your findings" or "based on the research." These phrases
delegate understanding to the worker instead of doing it yourself.

## Worker Prompt Best Practices

Workers cannot see your conversation. Every prompt must be self-contained:

**Good — synthesized spec:**
"Fix the null pointer in src/auth/validate.ts:42. The user field is undefined when
sessions expire but the token is still cached. Add a null check before user.id access.
Commit and report the hash."

**Bad — lazy delegation:**
"Based on your findings, fix the auth bug."
"The worker found an issue. Please fix it."

### Purpose Statement

Include a brief purpose so workers calibrate depth:

- "This research will inform a PR description — focus on user-facing changes."
- "I need this to plan an implementation — report file paths, line numbers, and type signatures."
- "This is a quick check before we merge — just verify the happy path."

## Continue vs Spawn Decision

| Situation                                             | Action                   | Why                                |
| ----------------------------------------------------- | ------------------------ | ---------------------------------- |
| Research explored exactly the files that need editing | Continue existing worker | Worker has files in context        |
| Research was broad, implementation is narrow          | Spawn fresh worker       | Avoid context noise                |
| Correcting a failure                                  | Continue existing worker | Worker has error context           |
| Verifying code a different worker wrote               | Spawn fresh worker       | Fresh eyes, no implementation bias |
| Wrong approach entirely                               | Spawn fresh worker       | Clean slate avoids anchoring       |

## Parallel Execution

**Parallelism is the primary advantage.** Launch independent workers concurrently:

- Read-only tasks (research) — run in parallel freely
- Write-heavy tasks (implementation) — one at a time per set of files
- Verification can sometimes run alongside implementation on different file areas

## Scratchpad Pattern

For durable cross-worker knowledge, use a shared scratchpad file:

- Workers write findings to a known location
- Other workers read from it when they need context
- Structure files by topic, not by worker

## Multi-Terminal Workflow

When running multiple Claude Code terminals (Boris Mode):

- Each terminal works on a separate feature branch or file area
- Use /sync on develop after merging to reconcile shared-state files
- Use /conflict-resolver if merges produce conflicts
- Shared-state files (PROGRESS.md, SPEC.md, package.json version) are ONLY modified on develop