feat(nano): operational constraint-checklist verification guideline (PR 19) - #898
Merged
Merged
Conversation
Four of the nine completed-wrong tasks in tb21-nano-flash-max-2 were one-assertion near-misses, each violating a constraint stated in the task text that the agent never tested: build-pov-ray built from the wrong source archive (2/3 tests green), dna-insert's primer Tm delta was 6.5°C against a stated ≤5°C bound, filter-js-from-html claimed 'fully verified' while mutating clean HTML entities, pytorch-model-cli passed 5/6 with an untested CLI output format. The bullet now names the method — extract the task's explicit requirements AND constraints (formats, tolerances, numeric bounds, required sources/methods) into a list and test the artifact against each — rather than only the principle. TB verifiers are hidden, so self-verification against the stated constraints is the only available check; this is the general skill, not benchmark knowledge. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
agentforce314
added a commit
that referenced
this pull request
Aug 16, 2026
…constraint-checklist"" This reverts commit 3e705e4.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Four of nine completed-wrong v2 tasks were one-assertion near-misses on constraints stated in the task (wrong source archive, Tm 6.5°C vs ≤5°C, entity mutation, untested CLI output). The verification bullet now prescribes the method: list the task's explicit requirements and constraints, test the artifact against each. General skill; +~25 est. tokens.
🤖 Generated with Claude Code