Cursor + Claude for whole-team coding (L2)

Standardize the team on Cursor or Windsurf with Claude Sonnet as the default model. Shared .cursorrules / .windsurfrules file in every repo so the AI follows your team conventions, not generic JavaScript advice.

WORKFLOW1Pick one IDE: Cursor or W…ndsurfCursor2Write .cursorrules per re…oCursor3Set Claude Sonnet as defa…ltClaude4Weekly tip-shareSlack
4 steps, in order, with the tool that owns each one.
Adoption ladderSix levels from Starter to Rebuilt. This item sits at level 2.L1 StarterOne tool, no workflow changeL2 AssistedAI drafts, humans approveL3 IntegratedWired into CRM and SlackL4 OrchestratedMulti-step, owned, measuredL5 AutonomousAgent runs, human auditsL6 RebuiltThe process itself changes
This playbook belongs at L2 Assisted. Running it above your level is how pilots stall.
Measures of successPR review cycle time; % of PRs with AI-assisted authoring; onboarding time for new hiresPROVE IT WORKEDPR review cycle time% of PRs with AI-assisted authoringonboarding time for new hires

The steps

  1. 01

    Pick one IDE: Cursor or Windsurf

    Tool: Cursor

    Cursor = larger community, more polished UX, $20/user/month. Windsurf = stronger "Cascade" agentic flows, comparable price. Pick one for the team, fragmenting will splinter your .cursorrules / .windsurfrules investment. Owner: Eng leadership. Time: 1 week of evaluation. Pitfall: "let engineers pick whichever", you lose the team-conventions network effect. DoD: every IC migrated within 2 weeks; old IDEs uninstalled or relegated to side projects.

  2. 02

    Write .cursorrules per repo

    Tool: Cursor

    In each repo's root, add a .cursorrules (or .windsurfrules) file with: stack, naming conventions, testing framework, what to never touch (generated files, migrations after merge), preferred patterns. Treat it like a CONTRIBUTING.md for AI. Example bullets: "Use Vitest, not Jest", "Prefer named exports", "All Supabase queries go through src/lib/db.ts". Owner: tech lead per repo. DoD: every actively developed repo has a .cursorrules file referenced in PRs.

  3. 03

    Set Claude Sonnet as default

    Tool: Claude

    In team Cursor settings, default to claude-sonnet-4 (or latest). Sonnet beats GPT-4 for codebase reasoning in most internal benchmarks, and Anthropic's data-handling terms are friendlier for source code than OpenAI consumer accounts. Owner: Eng ops. Pitfall: leaving free-tier models on, engineers will hit rate limits, get frustrated, and quietly stop using AI. DoD: usage dashboard shows >70% of completions from the chosen Sonnet model.

  4. 04

    Weekly tip-share

    Tool: Slack

    Create #ai-coding-tips Slack channel. Each Friday, one engineer posts the best Cursor prompt or workflow they used that week (e.g. "composed a multi-file refactor with @-mentioning these 5 files"). After 4 weeks you have a team wiki of real workflows, not vendor demos. Owner: rotating. DoD: channel has >4 weekly tip posts running.

Tools in this playbook

Next playbooks

Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.

Share this playbook

Posting to Instagram or TikTok? Copy the link, it carries the title, summary and share image.

Arrives weekly by email. Free. Unsubscribe anytime. By subscribing you agree to our Privacy policy and Terms. We never sell or share the list.