taskflow

Coordinate detached multi-step jobs with durable flow identity and revision-checked state.

Updated Apr 8, 2026
One-click install
npx skills add https://github.com/JOHNNYWHITEMIKE/openclaw --skill taskflow-johnnywhitemike
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: taskflow
Source: https://github.com/JOHNNYWHITEMIKE/openclaw/tree/main/openclaw/.openclaw/openclaw/skills/taskflow
Command: npx skills add https://github.com/JOHNNYWHITEMIKE/openclaw --skill taskflow-johnnywhitemike

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

TaskFlow solves the problem of coordinating multi-step background work across detached tasks while preserving a single owner context and a resumable execution identity.

Core Features & Use Cases

  • Durable flow identity for orchestration: Maintain one flow id that links child tasks and represents the overall job across runs.
  • Persisted state and waiting/resume semantics: Store orchestration state in stateJson, represent external waits with waitJson, and resume with revision-checked updates.
  • Revision-safe lifecycle control: Create, mutate, wait, resume, finish, fail, cancel, or request cancellation with conflict-safe mutation using flow.revision.

Quick Start

Use taskflow to create a managed flow, run child tasks linked to the same flow id, set a waiting state for an external reply, then resume and finish with the latest persisted stateJson.

Frequently Asked Questions about taskflow

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I resume detached background jobs after a system restart?

Detached background jobs resume after a restart by using a durable flow identity to link child tasks and persisting orchestration state in stateJson. This single flow identity preserves the owner context and allows execution to continue seamlessly across runs.

What is the best way to orchestrate multi-step workflows waiting on external systems?

Orchestrating multi-step workflows waiting on external systems requires setting a waiting state with waitJson to represent the external dependency. The workflow resumes with revision-checked updates once the external reply is received, ensuring conflict-safe mutations.

How do I prevent state conflicts when running parallel background tasks?

Preventing state conflicts when running parallel background tasks requires revision-safe lifecycle control using flow.revision. This mechanism applies conflict-checked mutations to safely create, wait, resume, finish, or cancel managed operations.

Can I link multiple child tasks under a single job identity?

Linking multiple child tasks under a single job identity is the core function of durable flow orchestration. You run child tasks linked to the same flow id, which represents the overall job and maintains one unified owner context across all detached executions.

What are the limitations of using detached tasks for long-running orchestration?

Detached tasks for long-running orchestration require managed lifecycle operations like createManaged, runTask, and resume to maintain state. Without persisted stateJson and waitJson, external waits or revision conflicts break the execution chain and lose progress.