agent-conduct

Enforces product isolation and peer quality check rules across squad agents.

Updated Jun 2, 2026
One-click install
npx skills add https://github.com/codebytes/btt --skill agent-conduct-codebytes
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-conduct
Source: https://github.com/codebytes/btt/tree/main/.squad/templates/skills/agent-conduct
Command: npx skills add https://github.com/codebytes/btt --skill agent-conduct-codebytes

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Squad agent charters previously duplicated the same two hard rules in every file, creating maintenance overhead and inconsistency risk. This Skill centralizes those shared governance rules so they are defined once and loaded by every agent. ## Core Features & Use Cases - Product Isolation Rule: Prevents tests, CI workflows, and product code from depending on specific squad agent names, requiring generic or parameterized values and obviously-fake test fixtures like "test-agent-1". - Peer Quality Check: Requires agents to run the test suite for touched files and verify changes do not worsen existing CI failures before declaring work done. - Use Case: A squad of 20 AI agents shares one governance skill instead of duplicating rules in each charter, so a policy update applies everywhere by editing a single file. ## Quick Start Load the agent-conduct skill and apply its product isolation and peer quality check rules to all work in this repository.

Frequently Asked Questions about agent-conduct

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

FAQPage Schema
How do I share common rules across multiple AI agent charters?

Extract duplicated rules into a single shared skill file with YAML frontmatter defining name and description. Each agent loads this skill once instead of repeating the rules in every charter, so updates propagate from one location.

What is the Product Isolation Rule for agent squads?

The Product Isolation Rule prohibits tests, CI workflows, and product code from depending on specific squad agent names. Use generic parameterized values or obviously-fake fixtures like "test-agent-1" instead of real agent names such as Flight or EECOM.

How should agents verify changes before finishing work?

The Peer Quality Check rule requires running the test suite for every file you touched and confirming your changes do not worsen pre-existing CI failures. Lessons learned from mistakes should be recorded in the agent's history.md file.

Can tests reference real squad agent names as fixtures?

No, tests must not hardcode real agent names from any squad. Use obviously-fake test fixtures such as "test-agent-1" or "TestBot" so product code and CI remain independent of any particular squad composition.

When does this governance skill not apply?

It applies only to squad agent conduct and shared team rules. It does not cover product feature logic, deployment configuration, or domain-specific coding standards, which belong in separate skills or charters.