โ† AI library

Editorial recipe โ€” not benchmarked โ€” reviewed September 3, 2026

Turn a software idea into a tested implementation

For teams that can state acceptance criteria and review agent-authored changes.

Stack
GitHub Spec Kit โ†’ Claude Code, Codex, or Qwen Code โ†’ Playwright โ†’ Chrome DevTools MCP

Run this recipe

Work through it here

Progress stays in this browser. The downloaded Markdown kit works in any notes app or repository.

0/5 complete

Is this recipe useful?

Procedure

  1. Pin the Spec Kit and agent versions, create a clean branch, and write the user outcome, exclusions, acceptance criteria, and rollback boundary before generating a plan.
  2. Review the generated specification, plan, and task list for unsupported assumptions; remove work that does not contribute to an acceptance criterion.
  3. Let one coding agent implement the approved tasks in an isolated worktree with only the repository permissions and test credentials it needs.
  4. Run existing checks plus the smallest Playwright user-flow proof; use Chrome DevTools MCP in an isolated profile to capture console, network, and performance evidence where relevant.
  5. Have a different model compare the final diff and evidence with the original specification, then resolve every material mismatch before merging.

Acceptance artifact

A linked specification, plan, task ledger, reviewed diff, executable checks, browser evidence, and rollback note

Do not use it blindly

Structured planning is not proof of correctness; never let generated tasks expand scope silently or expose authenticated browser data to an MCP client.

Evidence and setup