Bulk-refactor a 1M-line codebase for stricter compiler settings
Claude systematically converts whole classes of compiler errors into cleaner code, using custom skills or scripts, across 200K+ lines in a large codebase. ·
Done withClaude
- What they did
- On a mono-repo of about 50 projects and over 1M lines, they tighten compiler settings, which surfaces tens of thousands of errors. Claude takes whole classes of errors and converts them to cleaner code, using the team's custom skills or by writing new scripts. They sort refactors into logic, performance, and cosmetic buckets, and bulk-run only the latter two.
- How it went
- Claude handled changes across 200K+ lines that the author calls near impossible for a human in an IDE. The team also ported all code to dotnet 10 LTS and can upgrade dependencies wholesale. No timings are given.
- Worth knowing
- Keep business-logic changes out of bulk runs and in dedicated sessions, since the domain is complex and breakages are unaffordable; bulk-refactor only performance and cosmetic changes.
Try it yourself with Claude
In the codebase at [path], I want to enable the stricter compiler setting [flag or warning level]. Build the project, group the resulting errors by type, and pick the most common type first. Write a reusable script or skill that fixes that whole class of error safely, apply it, and rebuild to confirm the error count dropped and tests still pass. Repeat for the next error type, and stop when the build is clean or you've done [number] rounds. Show me the diff summary before you commit anything.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: HN · Nov 2, 2026
Related use cases
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Triage support emails and draft refund-policy replies
A consultant built a Claude agent that reads support Gmail, looks up orders, applies refund policy, drafts replies for human approval and creates triage records with optional Slack alerts.
Review AI-written code with a layered agent review process
Anthropic's Claude Code team describes how it changed its code review process so Claude can write most of Claude Code's code.
Turn tickets into merged pull requests
Pulls tasks from GitHub Issues or Linear, then runs Claude Code or Codex in isolated Kubernetes pods to carry each one through to a PR.
Build a budgeting app from a single prompt
A low-coder used Replit Agent to build a 'Budget Buddy' with category budgets, a spending dashboard and expense log in about 20 minutes, using half their monthly credits.
Run a task-manager sub-agent with atomic commits
A poster built a Claude Code sub-agent that fetches the next task, marks it complete, and enforces atomic commits with a mandatory commit id.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.