docs(plan): correct the drift-check note on eval case 1 - #4
Open
lozit wants to merge 1 commit into
Open
Conversation
…ld be permanent noise The two loop copies diverge on purpose (TODO.md vs loop/backlog.md, ADR 0030, plus checks the prototype never had), so the "cheap mechanical diff" noted alongside eval case 1 would be red forever. A useful check compares only the sections that must not drift; that is its own task, not a free diff.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
One sentence, caught verifying after PR #3 merged.
PLAN.mdcase 1 suggested pairing the eval case with "the cheap mechanical version — adiffof the two directories". That check would be red forever: the two loop copies diverge deliberately — the prototype drivesTODO.md, the shipped template drivesloop/backlog.md(ADR 0030), and the template carries Stage-1 checks the prototype never had. A permanently-red check is noise a real failure hides in, which is the opposite of the point.The note now says what a useful check would actually compare — only the sections that must not drift (the verifier's frame and what it is handed) — and that it is its own small task rather than a free
diff. The eval case itself is unchanged: it tests whether the agent reasons about the right layer, not whether the files match.