/orchestrate
Coordinates multiple specialized agents in parallel. Triggers: orchestrate, multi-agent, parallel agents, coordinate agents.
$ npx -y skills add softspark/ai-toolkit --skill orchestrate --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/orchestrate
Context preview
The summary Claude sees to decide when to auto-load this skill.
Coordinates multiple specialized agents in parallel. Triggers: orchestrate, multi-agent, parallel agents, coordinate agents.
SKILL.md
orchestrate.SKILL.mdname: orchestrate
description: "Coordinates multiple specialized agents in parallel. Triggers: orchestrate, multi-agent, parallel agents, coordinate agents."
user-invocable: true
effort: max
argument-hint: "[task]"
context: fork
agent: orchestrator
model: opus
allowed-tools: Bash, Read, Write, Edit, Glob, Grep, Agent, TeamCreate, TeamDelete, SendMessage, TaskCreate, TaskList, TaskUpdate, TaskGet, TaskOutput, TaskStop
/orchestrate - Multi-Agent Coordination
$ARGUMENTS
MANDATORY: You MUST use the Agent tool
**DO NOT analyze or implement this task yourself.** Your only job is to decompose the task and invoke specialized agents via the `Agent` tool. If you attempt to do the work inline, you have failed.
Step 1 — Decompose
Analyze the task and identify 3–6 sub-domains (e.g. backend, frontend, security, testing, docs). Define clear file ownership per domain so agents don't conflict.
Present the decomposition to the user and wait for approval before Step 2.
Step 2 — Spawn agents in parallel (REQUIRED)
Call the `Agent` tool **multiple times in a single response** — one call per sub-domain. All independent agents MUST be launched in the same message to run in parallel.
Example for a feature implementation task:
Agent(subagent_type="backend-specialist", prompt="...", ...)
Agent(subagent_type="frontend-specialist", prompt="...", ...)
Agent(subagent_type="test-engineer", prompt="...", ...)
Agent(subagent_type="security-auditor", prompt="...", ...)
Each agent prompt MUST include: 1. The original user task 2. The specific sub-task this agent owns 3. File paths this agent is allowed to modify 4. Success criteria for this agent's work
Step 3 — Synthesize
After all agents complete, generate the Orchestration Report combining their findings.
Available Agents
| Agent | Domain | |-------|--------| | `backend-specialist` | API, server logic, databases | | `frontend-specialist` | React, Vue, UI components | | `test-engineer` | Unit, integration, E2E tests | | `security-auditor` | OWASP, vulnerabilities, auth | | `database-architect` | Schema, migrations, queries | | `devops-implementer` | Docker, CI/CD, infra | | `performance-optimizer` | Profiling, bottlenecks | | `documenter` | KB, architecture notes, runbooks | | `code-reviewer` | Code quality, patterns | | `tech-lead` | Architecture, standards |
Output Format
## Orchestration Report
### Task
[Original task]
### Agents Invoked
| # | Agent | Sub-task | Files | Status |
|---|-------|----------|-------|--------|
| 1 | backend-specialist | API layer | src/api/ | Done |
| 2 | frontend-specialist | UI components | src/components/ | Done |
| 3 | test-engineer | Test suite | tests/ | Done |
### Key Findings
1. **[Agent]**: Finding
2. **[Agent]**: Finding
### Summary
[Synthesis]
Read more
name: orchestrate description: "Coordinates multiple specialized agents in parallel. Triggers: orchestrate, multi-agent, parallel agents, coordinate agents." user-invocable: true effort: max argument-hint: "[task]" context: fork agent: orchestrator model: opus allowed-tools: Bash, Read, Write, Edit, Glob, Grep, Agent, TeamCreate, TeamDelete, SendMessage, TaskCreate, TaskList, TaskUpdate, TaskGet, TaskOutput, TaskStop
/orchestrate - Multi-Agent Coordination
$ARGUMENTS
MANDATORY: You MUST use the Agent tool
**DO NOT analyze or implement this task yourself.** Your only job is to decompose the task and invoke specialized agents via the `Agent` tool. If you attempt to do the work inline, you have failed.
Step 1 — Decompose
Analyze the task and identify 3–6 sub-domains (e.g. backend, frontend, security, testing, docs). Define clear file ownership per domain so agents don't conflict.
Present the decomposition to the user and wait for approval before Step 2.
Step 2 — Spawn agents in parallel (REQUIRED)
Call the `Agent` tool **multiple times in a single response** — one call per sub-domain. All independent agents MUST be launched in the same message to run in parallel.
Example for a feature implementation task:
Agent(subagent_type="backend-specialist", prompt="...", ...) Agent(subagent_type="frontend-specialist", prompt="...", ...) Agent(subagent_type="test-engineer", prompt="...", ...) Agent(subagent_type="security-auditor", prompt="...", ...)
Each agent prompt MUST include: 1. The original user task 2. The specific sub-task this agent owns 3. File paths this agent is allowed to modify 4. Success criteria for this agent's work
Step 3 — Synthesize
After all agents complete, generate the Orchestration Report combining their findings.
Available Agents
| Agent | Domain | |-------|--------| | `backend-specialist` | API, server logic, databases | | `frontend-specialist` | React, Vue, UI components | | `test-engineer` | Unit, integration, E2E tests | | `security-auditor` | OWASP, vulnerabilities, auth | | `database-architect` | Schema, migrations, queries | | `devops-implementer` | Docker, CI/CD, infra | | `performance-optimizer` | Profiling, bottlenecks | | `documenter` | KB, architecture notes, runbooks | | `code-reviewer` | Code quality, patterns | | `tech-lead` | Architecture, standards |
Output Format
## Orchestration Report ### Task [Original task] ### Agents Invoked | # | Agent | Sub-task | Files | Status | |---|-------|----------|-------|--------| | 1 | backend-specialist | API layer | src/api/ | Done | | 2 | frontend-specialist | UI components | src/components/ | Done | | 3 | test-engineer | Test suite | tests/ | Done | ### Key Findings 1. **[Agent]**: Finding 2. **[Agent]**: Finding ### Summary [Synthesis]
Professional-grade AI coding toolkit with multi-platform support. Machine-enforced safety, 109 skills, 44 agents, expanded lifecycle hooks, persona presets, experimental opt-in plugin packs, and benchmark tooling — works with Claude Code, Claude Chat/Cowork,
Repo: softspark/ai-toolkit
Other skills on ai-toolkit.
- /ai-toolkit-rules
Mandatory engineering, security, testing, git, performance, quality, and response rules. Claude MUST load this skill for every technical, coding, debugging, review, architecture, DevOps, data, or file-editing task in Chat or Cowork.
Open skill - /mem-search
Search past coding sessions using natural language. Finds relevant observations, decisions, and context from previous work.
Open skill - /a11y-validate
Accessibility validator: WCAG 2.1 AA, EN 301 549, EAA. Triggers: a11y, accessibility, WCAG, EAA, ARIA, contrast, keyboard, screen reader.
Open skill - /agent-creator
Creates new specialized agents with frontmatter, tools, delegation. Triggers: new agent, create agent, agent scaffold, specialized agent.
Open skill - /analyze
Analyzes code quality, complexity, patterns across codebase. Triggers: quality report, hotspot scan, code analysis, architecture signal.
Open skill - /api-patterns
REST/GraphQL API design: naming, versioning, pagination, idempotency, OpenAPI. Triggers: API design, REST, GraphQL, OpenAPI, Swagger, idempotency, rate limit.
Open skill

