Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.stagewise.io/llms.txt

Use this file to discover all available pages before exploring further.

Skills and context files give the agent deeper understanding of your project. They are configured per workspace in Settings → Agent → Skills & Context files.

​
Context files

​
WORKSPACE.md

When you connect a workspace, stagewise can auto-generate a WORKSPACE.md file in the .stagewise/ directory. This file contains:
  • Project type, languages, and runtime
  • Package structure and dependency graph
  • Build scripts and CI configuration
  • Key files and their purposes
  • Architecture overview
The agent reads this file automatically to understand your project structure. stagewise also updates this file automatically when it detects significant changes to your project. Generating or updating WORKSPACE.md:
  1. Open Settings → Agent → Skills & Context files
  2. Select the workspace
  3. Click Generate (or Regenerate if one exists)
The file is written to .stagewise/WORKSPACE.md in your project. You can commit it to source control or add it to .gitignore.

​
AGENTS.md

An optional AGENTS.md file in your project root provides custom instructions to the agent. Use this for:
  • Coding conventions and style rules
  • Preferred libraries and patterns
  • Project-specific terminology
  • Build/test commands
The agent respects AGENTS.md when it’s present, but it’s opt-in — you enable it per workspace in the Skills & Context settings.
We recommend ignoring AGENTS.md as it tends to do more harm than good — the information can get outdated quickly and mislead the agent.

​
Skills

Skills are structured instruction sets that extend the agent’s capabilities for specific tasks. A skill is a folder with a SKILL.md file and optional supporting files.

​
Skill locations

Skills are loaded from two places (in priority order):
  1. .stagewise/skills/ — stagewise-exclusive skills (highest priority)
  2. .agents/skills/ — Skills shared across multiple AI tools

​
Skill structure

.stagewise/skills/my-skill/
├── SKILL.md          # Main instructions (required)
├── references/       # Reference files the agent can read
├── assets/           # Supporting assets
└── scripts/          # Script examples (read-only, not executed)

​
Creating a skill

A SKILL.md file needs a name and description in its frontmatter. Skills follow the Agent Skills Standard: 
---
name: "my-skill"
description: "When and how the agent should use this skill."
---

# Instructions

Your structured instructions here...
The description field tells the agent when to activate the skill. Be specific — e.g. “Use when creating or reviewing React components that use Zustand stores.”

​
Enabling skills per workspace

In Settings → Agent → Skills & Context files, you can toggle individual skills on or off for each workspace. Disabled skills are ignored by the agent even if present on disk.