"Check the environment before the prompt" — until the prompt actually was the bug
ClickOk5811 · reddit · 2026-10-17
The author has long argued that when an AI feature misbehaves, the prompt is the last place to look — check input schemas, upstream changes, tool descriptions, assembled context first. Then he hit a counterexample: same input, tools, context and model version, with a six-week-old untouched prompt that was still the culprit. It had never precisely handled one edge case; growing usage just made that edge case show up more often. The prompt was always wrong — volume caught up with it.
His revised takeaway: checking the environment isn't valuable because the environment is usually the problem; it's valuable because it's cheap and tells you where the problem isn't, isolating the prompt as the remaining variable.
His working heuristic: if failures correlate with a change, look around the prompt; if they correlate with volume, look at the prompt itself — though he admits it's not clean and may not generalize beyond his setup.
More from coding & agent
- Backend migration from Python to Go now takes one engineer just one week — rakyll · 2026-10-17
- Google's playbook for agent evals: four layers of checks beyond the vibe check — rseroter · 2026-10-17
- AgentForesight: a 7B online auditor that catches multi-agent failures mid-run — jiqizhixin · 2026-10-17
- Dev gives his agent an alarm tool to wake him during overnight /goal runs — zack_overflow · 2026-10-17
- Dev shares multi-agent orchestration recipe: 2-3 lead agents with many focused subagents — MoonL88537 · 2026-10-17
- Claude engineer runs 5 Claude Code agents in parallel terminals — now routine — adonis_singh · 2026-10-17