Don't Let Coding Agents Redefine 'Done': A 4-Question Review Guard
withmagi · reddit · 2026-10-19
When an agent writes both code and tests, a green suite can mean they drifted together. The author separates review into 4 questions: what was promised, what change implements it, what observation would expose a broken promise, and whether the change also weakened the check — the easiest to miss in a diff.
Worked example from Lamp, a status page at Just Every: the promise (if the DB is unreadable, serve the last saved copy labelled with its age, never an all-clear) is written down first; the test pulls a table out from under the running DB and asserts "All systems operational" never appears. Two guards keep question 4 honest: the build reads its own list of promises and fails if one lacks a test, and the release check diffs promises against every past release so any change must be recorded.
Practical takeaway: pick one failure you refuse to accept, write a tiny fixture that triggers it, and require a written reason whenever that expectation changes. Lamp is MIT-licensed on GitHub, with code written entirely by coding agents.
More from coding & agent
- Dev ships a playable Tesla Roadster game in a weekend using Opus 5.5 — edwinarbus · 2026-10-19
- Musk amplifies Grok Bot feature list: own computer, email, and a team of bots — elonmusk · 2026-10-19
- HIVE team on agent access layers: approvals must bind exact arguments and policy versions — usehive · 2026-10-19
- PoA: An Experimental Network Where AI Agents Earn Coins by Writing Original Work — adamamcbride · 2026-10-19
- Next Token podcast: 'If AI gets something to 60 points, it's worth zero' — op7418 · 2026-10-19
- Running GLM on 4x DGX Spark locally to build a 3D Three.js portfolio site — 9r4n4y · 2026-10-19