start

Loads Trellis workflows and Git status to initialize AI-assisted coding sessions.

3|Updated Jan 25, 2026
One-click install
npx skills add https://github.com/BaSui01/agentflow --skill start-basui01
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: start
Source: https://github.com/BaSui01/agentflow/tree/main/.agents/skills/start
Command: npx skills add https://github.com/BaSui01/agentflow --skill start-basui01

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Quickly re-establishes project context and developer state at the start or resumption of a coding session so agents do not waste time rediscovering workflow rules, developer identity, git status, or active tasks. It prevents inconsistent or ad-hoc edits by ensuring work begins with the correct specs, injected context, and a clear plan.

Core Features & Use Cases

  • Workspace bootstrapping: Reads workflow guides and spec indexes from .trellis to learn repository standards before writing code.
  • Context gathering: Runs helper scripts to collect developer identity, git status, current and active tasks, and package spec indexes.
  • Task classification & routing: Classifies incoming requests into question, trivial fix, simple task, or complex task and routes them to direct edit, task workflow, or brainstorm workflows as appropriate.
  • End-to-end task workflow initiation: Creates task directories, writes PRDs, injects code-spec context into jsonl files, activates tasks, and orchestrates implement/check phases.
  • Use case: Begin a new feature or resume work on a paused ticket with guaranteed adherence to project conventions and automated context setup.

Quick Start

Begin by invoking the start skill to load .trellis workflow and spec indices, gather current task context, and ask what you would like to work on.

Frequently Asked Questions about start

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

FAQPage Schema
How do I initialize an AI coding session with existing project context?

To initialize an AI coding session, the start skill loads workflow guides, developer identity, git status, and active tasks from .trellis files. It ensures agents immediately understand repository standards and current work state without manual context gathering.

What is the best way to route incoming development tasks automatically?

Routing incoming development tasks is handled by classifying requests into questions, trivial fixes, simple tasks, or complex tasks. The start skill then directs work to direct edit, task workflow, or brainstorm workflows based on the assessed complexity.

How do I resume a paused coding task without losing workflow context?

Resuming a paused coding task requires loading previous session state and project guidelines. The start skill reads .trellis spec indexes, gathers current task context, and applies developer identity to ensure work resumes with correct conventions and injected context.

Does AI task workflow initiation require repository access to helper scripts?

AI task workflow initiation requires repository access to .trellis files and the ability to execute trellis helper scripts. These scripts create task directories, configure jsonl context, activate tasks, and run implement and check phases for end-to-end workflow management.

Why do AI agents make inconsistent edits when starting a new coding session?

Inconsistent edits happen when agents lack loaded workflow rules, developer identity, and project guidelines. The start skill prevents ad-hoc changes by bootstrapping workspace context from .trellis files before any code is written, ensuring adherence to repository standards.

Can I inject code-spec context into jsonl files for complex task workflows?

Injecting code-spec context into jsonl files is supported during end-to-end task workflow initiation. The start skill creates task directories, writes PRDs, configures jsonl context, and orchestrates implement and check phases for complex task execution.