dynamic-workflow

Decompose complex development objectives into structured subagent workflows with independent verification.

13|5|Updated Feb 25, 2026
One-click install
npx skills add https://github.com/Jonghakseo/my-pi --skill dynamic-workflow-jonghakseo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dynamic-workflow
Source: https://github.com/Jonghakseo/my-pi/tree/main/skills/dynamic-workflow
Command: npx skills add https://github.com/Jonghakseo/my-pi --skill dynamic-workflow-jonghakseo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents complex, long-running, or parallel development tasks from drifting, becoming inconsistent, or relying on unverified results by structuring them into coordinated agent workflows.

Core Features & Use Cases

  • Dynamic Task Decomposition: Break large objectives into focused work units with explicit scope, constraints, success criteria, and stop conditions.
  • Agent Orchestration: Select and coordinate specialized agents for implementation, research, review, security analysis, verification, and challenge testing.
  • Quality Gates and Recovery: Use parallel fan-out, sequential gates, independent verification, synthesis barriers, and bounded iteration loops to improve reliability.
  • Use Case: For a multi-file feature with security and architecture concerns, assign implementation to a worker, run independent verification and review, then synthesize findings and iterate only until defined quality criteria are met.

Quick Start

Use the dynamic-workflow skill to decompose the current complex task, coordinate the appropriate subagents, verify their results independently, and report remaining risks.

Frequently Asked Questions about dynamic-workflow

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

FAQPage Schema
How do I coordinate parallel subagents for complex multi-file development tasks?

Parallel subagents are coordinated by decomposing objectives into focused work units with explicit scope, routing specialized agents for implementation and review, and using synthesis barriers to independently validate results.

What is the best way to decompose uncertain development objectives into verified workflows?

Uncertain objectives are decomposed by defining explicit constraints, success criteria, and stop conditions for each work unit, then applying bounded iteration loops and quality gates to verify the synthesized results.

How do workflow verification gates prevent inconsistent results in agent orchestration?

Verification gates prevent inconsistent results by requiring independent validation of subagent outputs before synthesis, ensuring parallel fan-out tasks meet defined quality criteria before proceeding to the next phase.

When do I need bounded iteration loops for task decomposition in agent coordination?

Bounded iteration loops are needed when tasks are verification-sensitive or uncertain, allowing agents to iterate on findings only until defined quality criteria are met rather than running indefinitely.

Can I route specialized agents for security analysis and code review within the same workflow?

Specialized agents for security analysis and code review can be routed within the same workflow by assigning implementation to a worker agent, running independent verification, and synthesizing findings through sequential gates.

Why does complex parallel execution require explicit scope locking and approval controls?

Explicit scope locking and approval controls are required to prevent long-running parallel tasks from drifting, ensuring risky actions are approved and agents remain constrained to their assigned work units.