Cursor

Setup for Cursor

Cursor Quick Start

Install & Deploy

1. Install

npm install -g aiwg

2. Deploy to your project

cd /path/to/your/project

# Deploy all 4 artifact types for Cursor
aiwg use sdlc --provider cursor

3. Configure MCP (optional)

aiwg mcp install cursor

4. Open in Cursor

cursor .

5. Regenerate for intelligent integration

/aiwg-regenerate

`aiwg use` already wrote the project context, so natural-language command mapping ("run security review" → workflow) works right away via AIWG's Discover-First protocol. On an existing project, or to pull in the latest AIWG, run this to re-tailor the context to your codebase — optional on a brand-new project. See the Regenerate Guide for details.

6. You're ready. See the Intake Guide for starting projects.

What Gets Created

.cursor/
├── agents/      # SDLC agents (Requirements Analyst, Architecture Designer, etc.)
├── commands/    # Slash commands (/project-status, /security-gate, etc.)
├── skills/      # Skill directories with SKILL.md (voice profiles, project awareness, etc.)
├── rules/       # Context rules in MDC format (.mdc extension)
└── mcp.json     # MCP config (if enabled)

.aiwg/           # SDLC artifacts

Rules System

Cursor uses MDC format (`.mdc` extension) for rules with YAML frontmatter:

---
description: "Enforce token security practices"
globs: ["src/**/*.ts", "src/**/*.js"]
alwaysApply: false
---

# Token Security

Never hard-code tokens...

Rule Types

TypeConfigurationWhen It Activates
Always Apply`alwaysApply: true`Every session
Apply Intelligently`description` set, no `globs`Agent decides based on description
File-Scoped`globs` setWhen referenced files match pattern
ManualNo `alwaysApply`, no `globs`Only when user types `@rule-name`

.cursorrules vs .cursor/rules/

`.cursorrules` is a legacy root-level file. Cursor now recommends `.cursor/rules/` with MDC format. AIWG deploys rules to `.cursor/rules/` (native support) and optionally generates `.cursorrules` via the regenerate command for backward compatibility.

Using Agents

Invoke AIWG agents via @-mention in Cursor chat:

@security-architect Review the authentication implementation
@test-engineer Generate unit tests for the user service

Cursor 2.0+ supports automatic context gathering — the agent will search your codebase, read files, and run commands without needing explicit @-mentions for context.

Skills

AIWG deploys skills to `.cursor/skills/` in directory format (each skill has a `SKILL.md`). This aligns with Cursor's native skills system (introduced in 2.4). Skills are loaded on-demand when relevant, reducing context overhead.

Cloud Agents

Cursor's Cloud Agents (formerly Background Agents) run in isolated cloud VMs. They fully support MCP servers, making AIWG workflows accessible in cloud agent mode:

1. Configure AIWG MCP server in your Cloud Agent environment 2. Add AIWG instructions to `AGENTS.md` at repo root 3. Cloud Agents clone your repo, work on branches, and push changes

Cloud Agents can be triggered from Cursor Desktop, cursor.com/agents, Slack, GitHub, Linear, or API.

Multi-Agent / Worktrees

Cursor 2.0+ supports up to 8 agents running in parallel via Git worktrees. Configure via `Cursor Settings > Worktrees` or `.cursor/worktrees.json`.

Agent Loop

Agent loops support multi-provider execution. While Cursor agents are deployed via AIWG, agent task loops run through the CLI:

aiwg ralph "Fix all tests" --completion "npm test passes"

See Al Guide for full documentation including `--provider` options.

Troubleshooting

Natural language not working? Run regenerate:

/aiwg-regenerate

Rules not loading? Check file extension is `.mdc`, verify frontmatter syntax, and restart Cursor.

MCP tools not visible?

  • Verify `aiwg mcp serve` runs successfully standalone
  • Check `.cursor/mcp.json` syntax
  • View MCP logs: `Cmd+Shift+U` > select "MCP Logs"
  • Restart Cursor after config changes

Redeploy if needed:

aiwg use sdlc --provider cursor --force

Cursor Cloud Agents

Cursor Cloud Agents run in isolated cloud VMs and can be triggered from GitHub, Slack, Linear, and webhooks. AIWG works in Cloud Agent environments via the `install` step in `.cursor/environment.json`.

Create `.cursor/environment.json`:

{
  "install": "npm install -g aiwg && aiwg use sdlc --provider cursor",
  "terminal": {
    "env": {
      "AIWG_PROVIDER": "cursor",
      "AIWG_CONTEXT_WINDOW": "200000"
    }
  }
}

This tells the Cloud Agent VM to install AIWG and deploy the SDLC framework before starting work. A pre-configured template is available at:

agentic/code/frameworks/sdlc-complete/templates/cursor/environment.json.aiwg-template

MCP server in Cloud Agents:

The AIWG MCP server can also run inside the Cloud Agent VM. Add it to `.cursor/mcp.json`:

{
  "mcpServers": {
    "aiwg": {
      "command": "aiwg",
      "args": ["mcp", "serve"]
    }
  }
}

Automation triggers:

Cloud Agents can be triggered by commenting `@cursor` on GitHub issues or PRs, or via Slack/Linear. AIWG agent loops pair well with this: trigger a Cloud Agent on an issue, and Al handles the iterative fix cycle automatically.

Note: Test `npm install -g aiwg` in your target environment first — Cloud Agent VMs may have npm available but require `--prefix` or `--location=global` flags depending on the base image.

MCP Sidecar (Unrestricted AIWG Access)

For full unrestricted AIWG tool access (artifact management, workflow execution, template rendering), connect the AIWG MCP server as a sidecar:

aiwg mcp install cursor

See the Cursor MCP Sidecar Guide for complete setup including tool whitelisting and context optimization.