dan-workflow

Enforce the PLAN to APPLY to UNIFY to SUMMARY loop across DAN workflows.

Updated Mar 29, 2026
One-click install
npx skills add https://github.com/RavBogard/DAN --skill dan-workflow
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dan-workflow
Source: https://github.com/RavBogard/DAN/tree/main/.claude/skills/dan-workflow
Command: npx skills add https://github.com/RavBogard/DAN --skill dan-workflow

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Defines the shared protocols and conventions that ensure consistent, end-to-end DAN workflow orchestration across all agents and skills, enabling reliable planning, execution, and closed-loop delivery.

Core Features & Use Cases

  • PLAN -> APPLY -> UNIFY -> SUMMARY loop with explicit PLAN.md and SUMMARY.md relationships, ensuring no orphan plans and no skipping of the final SUMMARY phase.
  • Executor/Qualifier (E/Q) protocol with distinct roles and access: Executors generate artifacts with full tool access; Qualifiers read outputs for independent verification.
  • Diagnostic Routing to classify root causes (Intent, Spec, Code) before applying fixes, reducing wasted effort and guiding remediation.
  • State Protocol via dan-tools.cjs to read and write state centrally, including commands for state read/ get/ set/ patch and for parsing frontmatter.
  • Two-Level Rule: Skills orchestrate; Agents execute; prevents nested agent spawning for debuggability and predictable costs.
  • Fresh Context Rule: Each agent runs in a clean context with all needed inputs passed via file paths, and state persisted in STATE.md.

Quick Start

Provide a PLAN.md for a task and ensure a corresponding SUMMARY.md exists, then update STATE.md to close the loop.

Frequently Asked Questions about dan-workflow

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

FAQPage Schema
How do I orchestrate multi-agent workflows with strict state management?

You orchestrate multi-agent workflows by enforcing a PLAN -> APPLY -> UNIFY -> SUMMARY loop, using shared state tools to centrally read and write state for reliable execution and closed-loop task delivery.

What is the executor and qualifier protocol for task verification?

The executor and qualifier protocol separates roles so executors generate artifacts with full tool access, while qualifiers independently read those outputs to verify task results without modification privileges.

How do I prevent orphan plans when coordinating end-to-end task planning?

You prevent orphan plans by pairing every PLAN.md with a corresponding SUMMARY.md, ensuring the final summary phase is never skipped and closing the loop by updating STATE.md.

Can I run nested agents for complex diagnostic routing and task execution?

No, a two-level rule prevents nested agent spawning because skills orchestrate while agents execute, ensuring debuggability and predictable costs by running each agent in a fresh context.

How do I classify root causes before applying fixes to workflow tasks?

You classify root causes using diagnostic routing to categorize issues as Intent, Spec, or Code before applying fixes, which reduces wasted effort and guides targeted remediation.