Multi-Agent Workflow Guide
Scheduling
Goal
Guide manual multi-agent coordination for complex work that spans PM, frontend, backend, mobile, and QA responsibilities.
Intent signature
- User wants step-by-step coordination, manual agent spawning, or multi-domain work planning without full automation.
- Task spans multiple specialist agents and requires contract alignment.
When to use
- Complex feature spanning multiple domains (full-stack, mobile)
- Coordination needed between frontend, backend, mobile, and QA
- User wants step-by-step guidance for multi-agent coordination
When NOT to use
- Simple single-domain task -> use the specific agent directly
- User wants automated execution -> use orchestrator
- Quick bug fixes or minor changes
Expected inputs
- Complex feature or project goal
- Required domains and priority tiers
- Workspace/session constraints and API/data contract needs
Expected outputs
- Manual coordination sequence
- PM task decomposition, agent spawn order, monitoring guidance, and QA review step
- API/data contract alignment checkpoints
Dependencies
- PM, frontend, backend, mobile, QA, and orchestrator skills
- CLI
oma agent spawn and progress/result memory conventions
Control-flow features
- Branches by task complexity, priority tiers, dependency ordering, and whether automation is desired
- Spawns independent same-priority tasks in parallel when appropriate
- Monitors progress files and contract alignment
Structural Flow
Entry
- Confirm the task is complex enough for multi-agent coordination.
- Start with PM task decomposition.
- Identify priority tiers and shared contracts.
Scenes
- PREPARE: Define session, domains, and task decomposition needs.
- ACT: Spawn agents by priority with separate workspaces.
- VERIFY: Monitor progress and API/data contract alignment.
- FINALIZE: Run QA review and coordinate remediation.
Transitions
- If task is simple, route to one specialist.
- If user wants automated execution, use orchestrator.
- If QA finds CRITICAL issues, re-spawn responsible agents.
Failure and recovery
- If contracts diverge, pause downstream frontend/mobile work until backend/API contract is reconciled.
- If agent workspaces conflict, split ownership boundaries.
- If progress stalls, inspect progress files and reissue focused instructions.
Exit
- Success: specialist outputs are coordinated and QA-reviewed.
- Partial success: blocked agents, contract conflicts, or QA failures are explicit.
Logical Operations
Actions
| Action |
SSL primitive |
Evidence |
| Read request and domains |
READ |
User prompt and project context |
| Select agent plan |
SELECT |
PM decomposition and priority tiers |
| Spawn agents |
CALL_TOOL |
oma agent spawn |
| Monitor progress |
READ |
progress-{agent}[-{sessionId}].md |
| Validate contracts |
VALIDATE |
API/data model alignment |
| Notify coordination status |
NOTIFY |
Final coordination summary |
Tools and instruments
oma agent spawn, PM/frontend/backend/mobile/QA agents
- Memory/progress/result files
- Configured code intelligence, with native fallback per
../_shared/core/code-intelligence.md
Canonical command path
oma agent spawn pm <pm-prompt-file> <session-id> --task-id <plan.pm_task.id> -w ./pm
oma agent spawn backend <backend-prompt-file> <session-id> --task-id <plan.backend_task.id> -w ./backend &
oma agent spawn frontend <frontend-prompt-file> <session-id> --task-id <plan.frontend_task.id> -w ./frontend &
wait
When native runtime dispatch is available (per-agent target vendor equals the current runtime vendor), prefer the runtime's native subagent path and use oma agent spawn as the cross-vendor fallback — same resolution rule as oma-orchestration.
Useful agent spawn options: --vendor <vendor> (CLI vendor override), --isolation worktree (git worktree per spawn, prevents file conflicts), --read-only (non-destructive tools only, e.g. for review/QA passes).
Resource scope
| Scope |
Resource target |
LOCAL_FS |
Progress/result files and workspaces |
PROCESS |
Agent spawn commands |
MEMORY |
Session state and task board |
CODEBASE |
Shared contracts and implementation areas |
Preconditions
- Task requires multiple domains.
- PM decomposition can identify independent priority tiers.
Effects and side effects
- Spawns or guides multiple agents.
- Coordinates workspace ownership and QA feedback.
Guardrails
- Always start with PM Agent for task decomposition
- Spawn independent tasks in parallel (same priority tier)
- Define API contracts before frontend/mobile tasks
- QA review is always the final step
- Assign separate workspaces to avoid file conflicts (or use
--isolation worktree for a git worktree per spawn)
- Follow
../_shared/core/code-intelligence.md: discover configured tools, do not auto-install or track, and use native scoped search when unavailable or timed out
- Never skip steps in the workflow; follow each step sequentially without omission
Workflow
Step 1: Plan with PM Agent
PM Agent analyzes requirements, selects tech stack, creates task breakdown with priorities.
Step 2: Spawn Agents by Priority
Resolve the dispatch path per agent, then spawn:
- Resolve the per-agent target vendor from oma-config.yaml (
agents: override, else model_preset)
- If the target vendor equals the current runtime vendor and a native subagent path exists, use native dispatch
- Otherwise use
oma agent spawn for that agent
- Spawn all same-priority tasks in parallel using background processes
# Example: spawn backend and frontend in parallel
oma agent spawn backend backend-prompt.md session-id --task-id plan.backend_task.id -w ./backend &
oma agent spawn frontend frontend-prompt.md session-id --task-id plan.frontend_task.id -w ./frontend &
wait
Step 3: Monitor & Coordinate
- Use memory read tool to poll
progress-{agent}[-{sessionId}].md files (spawned agents write the session-suffixed form)
- Verify API contracts align between agents
- Ensure shared data models are consistent
Step 4: QA Review
Spawn QA Agent last to review all deliverables. Address CRITICAL issues by re-spawning agents.
Automated Alternative
For fully automated execution without manual spawning, use the orchestrator skill instead.
References
1---2name: oma-coordination-23description: "Coordinate assigned specialist tasks and handoffs manually. Use when supervising a multi-agent project step by step."4metadata:5 source: https://skillmd.com/skills/gabrielmoreira/oma-coordination-26---78# Multi-Agent Workflow Guide910## Scheduling1112### Goal13Guide manual multi-agent coordination for complex work that spans PM, frontend, backend, mobile, and QA responsibilities.1415### Intent signature16- User wants step-by-step coordination, manual agent spawning, or multi-domain work planning without full automation.17- Task spans multiple specialist agents and requires contract alignment.1819### When to use2021- Complex feature spanning multiple domains (full-stack, mobile)22- Coordination needed between frontend, backend, mobile, and QA23- User wants step-by-step guidance for multi-agent coordination2425### When NOT to use2627- Simple single-domain task -> use the specific agent directly28- User wants automated execution -> use orchestrator29- Quick bug fixes or minor changes3031### Expected inputs32- Complex feature or project goal33- Required domains and priority tiers34- Workspace/session constraints and API/data contract needs3536### Expected outputs37- Manual coordination sequence38- PM task decomposition, agent spawn order, monitoring guidance, and QA review step39- API/data contract alignment checkpoints4041### Dependencies42- PM, frontend, backend, mobile, QA, and orchestrator skills43- CLI `oma agent spawn` and progress/result memory conventions4445### Control-flow features46- Branches by task complexity, priority tiers, dependency ordering, and whether automation is desired47- Spawns independent same-priority tasks in parallel when appropriate48- Monitors progress files and contract alignment4950## Structural Flow5152### Entry531. Confirm the task is complex enough for multi-agent coordination.542. Start with PM task decomposition.553. Identify priority tiers and shared contracts.5657### Scenes581. **PREPARE**: Define session, domains, and task decomposition needs.592. **ACT**: Spawn agents by priority with separate workspaces.603. **VERIFY**: Monitor progress and API/data contract alignment.614. **FINALIZE**: Run QA review and coordinate remediation.6263### Transitions64- If task is simple, route to one specialist.65- If user wants automated execution, use orchestrator.66- If QA finds CRITICAL issues, re-spawn responsible agents.6768### Failure and recovery69- If contracts diverge, pause downstream frontend/mobile work until backend/API contract is reconciled.70- If agent workspaces conflict, split ownership boundaries.71- If progress stalls, inspect progress files and reissue focused instructions.7273### Exit74- Success: specialist outputs are coordinated and QA-reviewed.75- Partial success: blocked agents, contract conflicts, or QA failures are explicit.7677## Logical Operations7879### Actions80| Action | SSL primitive | Evidence |81|--------|---------------|----------|82| Read request and domains | `READ` | User prompt and project context |83| Select agent plan | `SELECT` | PM decomposition and priority tiers |84| Spawn agents | `CALL_TOOL` | `oma agent spawn` |85| Monitor progress | `READ` | `progress-{agent}[-{sessionId}].md` |86| Validate contracts | `VALIDATE` | API/data model alignment |87| Notify coordination status | `NOTIFY` | Final coordination summary |8889### Tools and instruments90- `oma agent spawn`, PM/frontend/backend/mobile/QA agents91- Memory/progress/result files92- Configured code intelligence, with native fallback per `../_shared/core/code-intelligence.md`9394### Canonical command path95```bash96oma agent spawn pm <pm-prompt-file> <session-id> --task-id <plan.pm_task.id> -w ./pm97oma agent spawn backend <backend-prompt-file> <session-id> --task-id <plan.backend_task.id> -w ./backend &98oma agent spawn frontend <frontend-prompt-file> <session-id> --task-id <plan.frontend_task.id> -w ./frontend &99wait100```101102When native runtime dispatch is available (per-agent target vendor equals the current runtime vendor), prefer the runtime's native subagent path and use `oma agent spawn` as the cross-vendor fallback — same resolution rule as oma-orchestration.103104Useful `agent spawn` options: `--vendor <vendor>` (CLI vendor override), `--isolation worktree` (git worktree per spawn, prevents file conflicts), `--read-only` (non-destructive tools only, e.g. for review/QA passes).105106### Resource scope107| Scope | Resource target |108|-------|-----------------|109| `LOCAL_FS` | Progress/result files and workspaces |110| `PROCESS` | Agent spawn commands |111| `MEMORY` | Session state and task board |112| `CODEBASE` | Shared contracts and implementation areas |113114### Preconditions115- Task requires multiple domains.116- PM decomposition can identify independent priority tiers.117118### Effects and side effects119- Spawns or guides multiple agents.120- Coordinates workspace ownership and QA feedback.121122### Guardrails1231241. Always start with PM Agent for task decomposition1252. Spawn independent tasks in parallel (same priority tier)1263. Define API contracts before frontend/mobile tasks1274. QA review is always the final step1285. Assign separate workspaces to avoid file conflicts (or use `--isolation worktree` for a git worktree per spawn)1296. Follow `../_shared/core/code-intelligence.md`: discover configured tools, do not auto-install or track, and use native scoped search when unavailable or timed out1307. Never skip steps in the workflow; follow each step sequentially without omission131132### Workflow133134#### Step 1: Plan with PM Agent135136PM Agent analyzes requirements, selects tech stack, creates task breakdown with priorities.137138#### Step 2: Spawn Agents by Priority139140Resolve the dispatch path per agent, then spawn:1411421. Resolve the per-agent target vendor from oma-config.yaml (`agents:` override, else `model_preset`)1432. If the target vendor equals the current runtime vendor and a native subagent path exists, use native dispatch1443. Otherwise use `oma agent spawn` for that agent1454. Spawn all same-priority tasks in parallel using background processes146147```bash148# Example: spawn backend and frontend in parallel149oma agent spawn backend backend-prompt.md session-id --task-id plan.backend_task.id -w ./backend &150oma agent spawn frontend frontend-prompt.md session-id --task-id plan.frontend_task.id -w ./frontend &151wait152```153154#### Step 3: Monitor & Coordinate155156- Use memory read tool to poll `progress-{agent}[-{sessionId}].md` files (spawned agents write the session-suffixed form)157- Verify API contracts align between agents158- Ensure shared data models are consistent159160#### Step 4: QA Review161162Spawn QA Agent last to review all deliverables. Address CRITICAL issues by re-spawning agents.163164### Automated Alternative165166For fully automated execution without manual spawning, use the **orchestrator** skill instead.167168## References