onboard

Onboard developers to the Trellis workflow with a three-part interactive curriculum.

1|Updated Apr 13, 2026
One-click install
npx skills add https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413 --skill onboard-sddkkk
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: onboard
Source: https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413/tree/main/.agents/skills/onboard
Command: npx skills add https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413 --skill onboard-sddkkk

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the friction of bringing new contributors up to speed on the Trellis AI-assisted workflow by teaching the system's core philosophy, how project-specific knowledge is injected, and how to prevent context drift while ensuring human oversight before commits.

Core Features & Use Cases

  • Three-part interactive curriculum: Part 1 explains core philosophy (AI memory, project-specific specs, context drift) and the system structure; Part 2 walks through five real-world workflows step-by-step; Part 3 guides customizing and verifying .trellis/spec/ guidelines.
  • Practical verification and handoff: Demonstrates when to run $start, $before-dev, $check-*, $finish-work, and $record-session, and emphasizes that humans validate and commit changes.
  • Use Cases: New developer onboarding, preparing project-specific guidelines for AI, debugging sessions, code review fixes, and large refactors that require cross-layer checks.

Quick Start

Ask the onboard skill to walk me through Trellis core concepts, demonstrate the five example workflows, and help customize or verify the .trellis/spec/ guidelines for this repository.

Frequently Asked Questions about onboard

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

FAQPage Schema
What is the Trellis AI-assisted workflow and how does it handle project-specific specs?

The Trellis AI-assisted workflow uses project-specific specs to restore AI memory and prevent context drift. It injects guidelines from `.trellis/spec/` files to ensure the AI operates with accurate, project-specific knowledge during development and refactoring.

How do I onboard new developers to an AI-assisted workflow with project-specific guidelines?

To onboard developers, use an interactive curriculum that teaches core philosophy, demonstrates real-world workflows, and guides the customization of `.trellis/spec/` guidelines. This ensures new contributors understand AI memory injection and context drift prevention.

When do I need to run Trellis workflow commands like $before-dev and $finish-work?

You run `$before-dev` before starting development, `$check-*` during code review fixes or refactors, and `$finish-work` or `$record-session` when ending a task. These commands verify context drift and restore AI memory for the next session.

Does the Trellis onboarding workflow support debugging sessions and large refactors?

Yes, the Trellis workflow supports debugging sessions and large refactors by guiding developers through cross-layer checks. It uses interactive examples to demonstrate how to maintain human oversight and validate changes before committing.

How do I customize and verify `.trellis/spec/` files for my repository?

You customize and verify `.trellis/spec/` files by following the third part of the interactive onboarding curriculum. This process guides you through populating project-specific guidelines and verifying context drift prevention for your repository.

Why does AI context drift happen during development and how is human oversight maintained?

AI context drift happens when the AI loses track of project-specific constraints during long sessions. Human oversight is maintained by requiring developers to validate changes and manually commit, using commands like `$check-*` to verify alignment before finishing.