---
title: "Comment by Adam Lewis on \"61.20% of Rejected Agent Replies Were Too Early or Too Much\" | Hashnode"
description: "Once you go past labelling failures, an agent acting unasked needs to earn its way up through the levels, so a mention that gets taken up is what buys the sugge"
url: "https://hashnode.com/posts/61-20-of-rejected-agent-replies-were-too-early-or-too-much/6ae9333c10224366ff8ce09d/comment/6aeda17fbc9327348393032f"
---

[Discussion](https://hashnode.com/posts/61-20-of-rejected-agent-replies-were-too-early-or-too-much/6ae9333c10224366ff8ce09d)

Comment [61.20% of Rejected Agent Replies Were Too Early or Too Much](https://hashnode.com/posts/61-20-of-rejected-agent-replies-were-too-early-or-too-much/6ae9333c10224366ff8ce09d)

[Adam Lewis](https://hashnode.com/@nark3d)

[6d ago](https://hashnode.com/posts/61-20-of-rejected-agent-replies-were-too-early-or-too-much/6ae9333c10224366ff8ce09d/comment/6aeda17fbc9327348393032f)

Once you go past labelling failures, an agent acting unasked needs to earn its way up through the levels, so a mention that gets taken up is what buys the suggestion, and only a suggestion that gets taken up buys an action. Too much too soon becomes the same agent having skipped a rung it wasn't credited for. A single pass-or-fail label on one response can't tell you that, because the failure lives between the responses. You'd have to score a run of them, letting it escalate only after the earlier, smaller move was accepted.

[prickles.org/tenet/single-responsibility-principl…](https://prickles.org/tenet/single-responsibility-principle/F1)

[kestrelfield](https://hashnode.com/@kestrelfield)

[4d ago](https://hashnode.com/posts/61-20-of-rejected-agent-replies-were-too-early-or-too-much/6ae9333c10224366ff8ce09d/comment/6aeda17fbc9327348393032f#reply-6af1237d78a2c2952bb0dd81)

That is a fair point, and it is the part my checklist does not cover. Scoring each response alone would miss an agent that jumps from a mention straight to an action. A simple way to try it is to log the sequence per scenario and only count an action as fine if a smaller step came first and was accepted. I would be curious how often that rule alone changes the pass rate.