Write exhaustive test suites to clear a coverage backlog
A team had Claude write exhaustive test suites for their coverage backlog, but the tests overloaded the CI node and covered disabled features, so it only partly worked. ·
Done withClaude
- What they did
- The team had Claude write exhaustive test suites to work through their code coverage backlog. The commenter doesn't describe the prompts or tooling in detail. After the tests landed in develop, the team noticed the suite had become flakier and slower, and a person traced the problem back to the generated tests.
- How it went
- The full suite normally ran 30-40 minutes but could exceed an hour and then fail. Claude also wrote tests for disabled features that were about to be rewritten, so the result was only partly useful.
- Worth knowing
- Tell the model what to skip and how much compute the tests may use, since 'exhaustive' got read literally and it covered disabled features slated for rewrite.
Try it yourself with Claude
In my repo at [path or repo name], find the files with the lowest test coverage in [module or directory] and write tests for them using [test framework]. Skip features that are disabled or behind flags (check [config file or flag list]), keep the suite light enough to run on a small CI node by avoiding heavy fixtures and long loops, and run the tests locally to confirm they pass and report the coverage change. Work in batches of [number] files and show me the results after each batch before moving on.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: HN · Nov 9, 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.
Survey AI product names with a swarm of search agents
The author had a swarm of Claude agents run 4,654 searches to catalog AI bot names and analyze naming trends by gender, industry and year.
Grade student submissions from Canvas against a rubric
An instructor's pipeline pulls student files from Canvas, has Claude score each per criterion with comments, and posts grades back after human review.
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.
Scaffold a new service end to end with agents
Cursor agents create Terraform, k8s manifests and CI, validate them and open a draft PR; one cleanup run wrongly deleted Terraform files.
Run Claude Code in an isolated VM that opens PRs
The author runs Claude Code in a restricted VM with secrets served from Infisical. It can only create issues and PRs on Forgejo, and merges are manual before CI deploys.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.