skills.rest
SDDKKK
SDDKKKCOMMUNITY
Explore all skills by this author →
GitHub

trellis-meta

Document Trellis architecture, hooks, agents, and customization procedures.

1|Updated Apr 13, 2026
Search Tags:#workflow#hooks#agents#worktree#claude-code#specs#trellis
One-click install
npx skills add https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413 --skill trellis-meta-sddkkk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: trellis-meta
Source: https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413/tree/main/.claude/skills/trellis-meta
Command: npx skills add https://github.com/SDDKKK/Trellis_Hiskens_backup_20260413 --skill trellis-meta-sddkkk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Provides a single authoritative reference for the vanilla Mindfold Trellis system so teams can understand its architecture, lifecycle hooks, agent pipeline, and multi-session worktree model without accidentally introducing project-specific changes into the canonical documentation.

Core Features & Use Cases

  • Architecture & Design: Detailed system layers (skills, hooks, persistence) and diagrams to explain how Trellis orchestrates AI agents and file-based context.
  • Hooks & Context Injection: Explanations of SessionStart, PreToolUse, and SubagentStop hooks and how they inject specs and enforce the Ralph Loop verification.
  • Agents & Pipelines: Agent definitions, allowed tools, JSONL context patterns, and multi-agent vs multi-session workflows for dispatch, implement, check, debug, research, and plan agents.
  • Multi-Session / Worktrees: Guidance and scripts for creating isolated git worktrees, starting sessions, monitoring, and creating PRs.
  • Customization & Upgrade Protocols: How-to guides for adding commands, agents, specs, and hooks plus a self-iteration protocol that mandates project-local documentation in trellis-local.
  • Use Cases: Onboarding new contributors, adding a custom agent, modifying hook behavior safely, or upgrading Trellis while preserving project customizations.

Quick Start

Read this trellis-meta skill to learn the canonical Trellis design and then create a project-local .claude/skills/trellis-local/SKILL.md to record any project-specific changes before modifying hooks, agents, or specs.

Frequently Asked Questions about trellis-meta

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

FAQPage Schema
How do I customize Claude Code agents and hooks for my project workflow?▼

To customize Claude Code agents and hooks, you create a project-local `.claude/skills/trellis-local/SKILL.md` file to record modifications. This approach safely injects project-specific context and enforces quality checks without altering canonical documentation.

How does Trellis inject context and enforce quality checks during AI agent sessions?▼

Trellis injects context and enforces quality checks using SessionStart, PreToolUse, and SubagentStop hooks. These lifecycle hooks automatically load JSONL context patterns and execute the Ralph Loop verification process during agent pipelines.

Can I run multiple isolated AI agent sessions using git worktrees?▼

Yes, you can run multiple isolated AI agent sessions using git worktrees. Trellis provides guidance and scripts for creating isolated worktrees, starting parallel sessions, monitoring progress, and creating pull requests independently.

How do I upgrade Trellis without losing my custom agents and commands?▼

To upgrade Trellis without losing customizations, follow the self-iteration protocol that mandates storing project-local documentation in `trellis-local`. This separates your custom commands and agents from the canonical system during upgrades.

What is the Ralph Loop verification process in AI workflow hooks?▼

The Ralph Loop verification process is a quality enforcement mechanism triggered by the SubagentStop hook. It ensures AI agents complete required validation checks before proceeding in the dispatch, implement, check, and debug pipeline.

Does Trellis work with Cursor for multi-agent pipeline orchestration?▼

Yes, Trellis works with Cursor and Claude Code for multi-agent pipeline orchestration. It provides agent definitions, allowed tool configurations, and JSONL context patterns to manage dispatch, implement, research, and plan agents across sessions.

Related Skills

Explore More Software Engineering Skills→
jkomorosjkomoros
9

session-startup

Initialize Claude Code sessions by fetching updates, loading workspace config, and starting servers.

Community
Basic
cryborg1211cryborg1211

vc:merge-worktree

Merge a git worktree branch into the main checkout and prune it.

Community
Intermediate
OriginHubAIOriginHubAI
1

start

Launch AI development sessions with phase-based task workflows.

Official
Advanced
oabdelmaksoudoabdelmaksoud
3

autonomous-loops

Provide patterns for autonomous AI agent loops from pipelines to RFC-driven DAGs.

Community
Advanced