discipline

Embed discipline rules into multi-agent pipeline phases with standardized metadata.

6|3|Updated Feb 11, 2026

Search Tags:#verification #multi-agent #discipline #evidence #spec #proof

Ready to run

npx skills add https://github.com/vinhnxv/rune --skill discipline-vinhnxv

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: discipline
Source: https://github.com/vinhnxv/rune/tree/main/plugins/rune/skills/discipline
Command: npx skills add https://github.com/vinhnxv/rune --skill discipline-vinhnxv

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Discipline Engineering provides a verifiable, proof-based framework to ensure every task in a multi-agent system is decomposed, understood, verified, enforced, and accountable to the original specification.

Core Features & Use Cases

Five-layer discipline model: Decomposition, Comprehension, Verification, Enforcement, and Accountability.
Enforces machine-verifiable proofs, anti-rationalization, and spec-continuity across pipelines.
Supports on-demand references (references/), assets, and potential scripts to guide agents.

Quick Start

Load the discipline SKILL to auto-load proof rules and enforce spec-compliance across agent tasks.

Frequently Asked Questions about discipline

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

FAQPage Schema

How do I enforce spec compliance in multi-agent pipelines?▼

To enforce spec compliance in multi-agent pipelines, you need a proof-based discipline framework that embeds verification rules into every task phase. This ensures agents remain accountable to the original specification through machine-verifiable proofs.

What is multi-agent verification and how does it work?▼

Multi-agent verification is a process that ensures tasks are decomposed, understood, and verified against specifications. It works by applying a five-layer discipline model—Decomposition, Comprehension, Verification, Enforcement, and Accountability—across agent teams.

How do I prevent agents from rationalizing outputs during task decomposition?▼

To prevent agents from rationalizing outputs, you apply anti-rationalization rules alongside spec-continuity checks during task decomposition. This enforces machine-verifiable proofs, holding every agent accountable to the original pipeline specification.

Can I use proof-based discipline to orchestrate agent tasks across different phases?▼

Yes, proof-based discipline orchestrates spec-compliant agent tasks across design, decomposition, verification, and enforcement phases. It embeds discipline rules into every stage to ensure standardized metadata outputs and pipeline accountability.

What are the limitations of applying discipline rules to complex agent pipelines?▼

Limitations of applying discipline rules include the overhead of maintaining machine-verifiable proofs and spec-continuity across highly complex pipelines. It requires strict adherence to the five-layer model to prevent enforcement gaps.

Do I need a specific framework to build spec-compliant multi-agent systems?▼

Building spec-compliant multi-agent systems requires a framework that supports a five-layer discipline model and machine-verifiable proofs. This ensures tasks are properly decomposed, verified, and enforced against the original specification.