trellis-start

Initializes Trellis development sessions by loading context, tasks, and workflow routing rules.

1|Updated Mar 19, 2026
One-click install
npx skills add https://github.com/hejiajiudeeyu/delegated-execution-workspace --skill trellis-start-hejiajiudeeyu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: trellis-start
Source: https://github.com/hejiajiudeeyu/delegated-execution-workspace/tree/main/.agents/skills/trellis-start
Command: npx skills add https://github.com/hejiajiudeeyu/delegated-execution-workspace --skill trellis-start-hejiajiudeeyu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Starting a coding session without project context leads to missed tasks, ignored guidelines, and wrong workflow choices. This Skill restores full session context by reading developer identity, git status, active tasks, and project guidelines from the .trellis/ directory, then routes the incoming request to the correct workflow. ## Core Features & Use Cases - Session Context Loading: Runs get_context.py scripts to surface identity, git status, current task, active tasks, and journal location. - Workflow Routing: Classifies incoming work and routes to brainstorm, direct edit, or task workflows based on whether an active task and prd.md exist. - Guideline Discovery: Reads package and spec layer index files so coding guidelines are loaded on demand before implementation begins. - Use Case: A developer opens a repository mid-project and asks the AI to continue work; the Skill detects the active task with an existing prd.md and loads Phase 2 step 2.1 instructions instead of starting from scratch. ## Quick Start Start my Trellis development session and tell me what task I should work on next.

Frequently Asked Questions about trellis-start

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

FAQPage Schema
How do I restore project context when starting a new AI coding session?▼

Run the session initialization steps that execute get_context.py to load developer identity, git status, current task, and active tasks from the .trellis directory. Then read the relevant spec index files to discover which guideline documents apply before coding.

How does the skill decide between brainstorm, direct edit, and task workflows?▼

It checks the current task state: an active task with prd.md routes to Phase 2 step 2.1, an active task without prd.md routes to the brainstorm skill, and no active task routes to brainstorm for multi-step work or direct answers for trivial edits.

What is the .trellis directory and what does it contain?▼

The .trellis directory stores session and project management data including scripts like get_context.py, a workflow.md guide, task directories with prd.md files, journals, and spec guideline indexes organized by package and layer.

When should I skip session initialization and just answer directly?▼

Skip the full workflow only for simple one-off questions or trivial edits where no task tracking is needed. Multi-step work or feature requests should always go through task classification and the brainstorm skill first.

What should I do when a Trellis update notification appears?▼

Copy the full 'Trellis update available:' line verbatim when summarizing session context. Do not shorten the operational command hints included in that line.