AI coding agents keep inflating test directories — and devs mostly ignore it
granawkins · x · 2026-09-07
Developer Grant Hawkins asks whether everyone just lets Codex/Claude build an 'exponentially-inflating test directory': agents keep piling up tests, and while he scans edits in PRs, he mostly ignores the test folder — a relatable quirk of AI-assisted coding workflows.
More from coding & agent
- InboxMinder: open-source Mac agent scores every Gmail, only 6 of 127 emails need you — ayushtweetshere · 2026-09-07
- Domoticz MCP Server lets AI assistants control smart home devices via MCP — modelcontextprotocol · 2026-09-07
- Vocab Voyage ships 20 MCP tools and 17 widgets for SAT/GRE vocab prep — modelcontextprotocol · 2026-09-07
- Every company is now building dev tools, because models made everyone a developer — jobergum · 2026-09-07
- When an agent run is wrong with no known-good baseline, what do you debug against? — Sensitive-Parsnip-12 · 2026-09-07
- CuePrecise: an open-source local MCP that turns long YouTube videos into searchable timelines — Forward-Associate523 · 2026-09-07