Player profile
AIWorks
An MIT-licensed supervisor for one Windows PC that runs Claude Code, Codex, and Cursor through an approved plan in isolated git worktrees, has a separate session verify each step, and delivers pull requests to GitHub.
Profile
Where it fits
AIWorks is a Node.js supervisor that runs on the owner's computer. A plan is a private JSON file with an explicit authorization; the supervisor hands each step to a worker CLI (Claude Code, Codex, or Cursor) inside its own git worktree, has a separate verification session check the result against the step's acceptance criteria, and opens a pull request on GitHub. It keeps going through worker crashes, restarts, and provider usage limits, and asks the owner only when a decision is needed.
It belongs in Parallel Coding-Agent Runners because the plan, not a chat session, is what gets scheduled: steps run as independent worker sessions in isolated worktrees, implementation and verification are separate sessions, and a local dashboard shows what is running, what needs a decision, and what was delivered. Its README reports that in its first day of operation it produced 63 checked pull requests and wrote 53% of its own code.
Builder web analytics
Measure outcomes from AIWorks-delivered changes
AIWorks delivers reviewed pull requests; it does not observe what a merged change does for users. Separately configured product analytics can close that loop.
This is an optional workflow, not a native AIWorks integration with Agent Analytics. Configure tracking on the deployed product and analytics access separately; this listing does not install tracking or report AIWorks runs to Agent Analytics.
First loop to measure
- a plan step is implemented by a worker session and verified by a separate session
- the reviewed pull request is merged and deployed to a separately instrumented product surface
- a human or an agent with separately configured analytics access reads the observed results
- the owner uses that evidence to write the next plan
Copyable prompt
If tracking and analytics access are already configured for the deployed change, report measured user outcomes for the observation window with their limitations. Otherwise, identify what is missing. Do not assume a native AIWorks integration or infer outcomes from pull request activity.