Repository navigation
📝 docs: plan: map each check to an owner and a level, as COR-104 - #429
Merged
Merged
Conversation
…r own file architecture.md holds 9,943 bytes, and just limits caps a growing file at 10,000. The list of the folders outside the code map now has its own file, with no change of its text, so that the next commit can add a section. architecture.md links the new file in place of the list.
D068 extends D067 with the owner and level map of the verification loop: core, the template, and each extension, at the levels of code, checks, prose, then QA and UAT. It says how a correction moves up a level and across owners, and it replaces the score of the 36 trial cases with three measures: the prose-only rules of each rule index, the level and owner of each correction in the ledger, and the real use of the loop by video-convert. architecture.md gets a short section that links D067 and D068, and the glossary defines a verification level apart from a layer. COR-104 holds ten cards, core and template first, then video-convert, which adopts the loop as #427 asks. It follows COR-102 in the queue. The evidence is in nicely-research, under synthesis/verification-map.
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the
 |
… history of old records Video-convert updates to a Nicely release, but D043 releases an item only when it ships. The last merge of T1 to T9 now releases a minor version, as this plan asks. An extension with Markdown build records starts a new build when it updates and keeps them as history. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
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.
Pascal wanted to know where the line sits between Nicely's core, the extension template, and one extension such as video-convert, and how to push the most verification into the core and the template. COR-103 placed each check, but nothing mapped the whole loop: code, checks and CI, prose and skills, then QA and UAT, for each owner. The six trial rounds also showed that their score stopped measuring: builds scored 36 of 36 while real defects sat outside the cases.
This pull request writes that map as D068, which extends D067 without moving its line, and plans it as COR-104 with 10 cards. The core and the template get each practice first. Video-convert adopts the loop last, through #427.
docs/qa/ledger.md, and real use by video-convertarchitecture.mdgets a short Verification section. Its "other folders" list moves to its own file, unchanged, because the file was at 9,943 of 10,000 bytesjust nextstill serves COR-102-T1No card is implemented here. The evidence is in pascalandy/nicely-research#23, which merges first so that the links resolve.
How it was made: scouts and analyses on 237 correction commits, two prototypes (an SDK analyzer that reaches an extension at its pinned version, and the TLC runner in the kit with a model of the finish gate), then an arena of two plans. A blind judge scored the base 23 and the other 20. Five ideas were grafted from the other plan.
Proof:
just checkpasses atcc560495. The review fixes change only the COR-104 file;go -C tools test ./cmd/planandjust limitspass on them, andjust signoffruns on the final head.Reviewed with codex-loop: 2 rounds with gpt-6-astra on high. Sol verdict: approved
by Opus-5.5 via Claude Code