workflow-orchestrator

Visual workflow builder for orchestrating multi-agent systems with branching, parallel execution, error handling, and retry logic. Use when building agent pipelines, DAG-based task orchestration, multi-agent coordination, or complex automation workflows. Covers step definition, conditional branching, fan-out/fan-in, failure recovery, and live monitoring.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "workflow-orchestrator" with this command: npx skills add evezart/workflow-orchestrator

Workflow Orchestrator

Build and run multi-agent workflows with DAG execution, branching, and error handling.

Quick Start

from orchestrator import Workflow, Step, Branch, Runner

wf = Workflow("data-pipeline")
wf.add_step(Step("fetch", agent="scraper", action="fetch_url", params={"url": "https://example.com"}))
wf.add_step(Step("extract", agent="parser", action="extract_text", depends_on=["fetch"]))
wf.add_step(Step("summarize", agent="writer", action="summarize", depends_on=["extract"]))
wf.add_step(Step("translate", agent="writer", action="translate", depends_on=["extract"]))
wf.add_step(Step("publish", agent="publisher", action="send", depends_on=["summarize", "translate"]))

runner = Runner()
result = runner.execute(wf)

DAG Execution Model

fetch → extract → summarize → publish
                 → translate ↗

Steps run in parallel when their dependencies are met. The publish step waits for both summarize and translate.

Step Definition

Step(
    name="unique_step_name",
    agent="agent_id",          # Which agent executes this
    action="tool_name",        # What action to perform
    params={},                 # Input parameters
    depends_on=[],             # Wait for these steps first
    retry=3,                   # Max retries on failure
    timeout_sec=300,           # Step timeout
    on_failure="skip",         # "skip", "retry", "abort", "fallback"
    fallback_step="plan_b",    # Run this step on failure
    condition="$.fetch.status == 200",  # Conditional execution
)

Features

  • Parallel execution: Steps with satisfied dependencies run concurrently
  • Conditional branching: JSONPath conditions determine which branches execute
  • Retry with backoff: Configurable retry count and exponential backoff
  • Timeout handling: Steps that exceed timeout are killed and handled per on_failure
  • Fallback steps: Alternative steps run when the primary fails
  • Live status: Query workflow state at any point during execution
  • Error propagation: Configure whether failures bubble up or are contained

Monitoring

status = runner.status(workflow_id)
# {"running": 2, "completed": 3, "failed": 0, "pending": 1}

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

V19 Trust Manifesto

Agent Community认知治理协议公开受信声明v1.5.0。V8.6 Agent OS三大终极协议(ITE意图交易引擎三阶段翻译/ASM环境状态监控EventBus自动响应/Dual-Track Consensus双轨共识ConflictSet驱动进化)+6个学术框架对齐(MIA/AIGA/GCL/SCF...

Registry SourceRecently Updated
Automation

NEXO Brain

Cognitive memory system for AI agents — Atkinson-Shiffrin memory model, semantic RAG, trust scoring, and metacognitive error prevention. Gives your agent per...

Registry SourceRecently Updated
Automation

Growth Engineer

Growth Engineer for mobile apps and agent runtimes including OpenClaw and Hermes. Correlate analytics, crashes, billing, feedback, store signals, and repo co...

Registry SourceRecently Updated
7110Profile unavailable
Automation

Agent News

Query verified AI agent news with citations, confidence scores, and Ethics Engine ratings — sourced, not generated. Use instead of generic web search for any...

Registry SourceRecently Updated
1423Profile unavailable