Rewrite the README around the reader's first hour - #34
Merged
Conversation
Reorder for a first-time evaluator: the loop, a real failing check, the boundaries before adoption (REQ-008.1.2), install/use, the anti-cheat mechanics with a real committed verdict, positioning vs. spec-generation tools and mutation testing, and why we built it. Move the file-scoped ID deep-dive to docs/file-scoped-ids.md and link panopticon as a live adoption example. Name the independent-runner pattern in Risks and replace the unreproducible 0.2s figure with a measured bound. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XUFtq5xDd1rvWbW6Vqu5Ws
tylerwillis
force-pushed
the
readme-rewrite
branch
from
August 8, 2026 07:06
3566a43 to
7c0650c
Compare
…scoped verdicts The condensed file-scoped-IDs summary dropped content REQ-011.8 requires the README itself to carry (both grammars with canonical-ID example, the 2119-spec sugar, and the rename-invalidates-verdicts paragraph). Expand the summary to satisfy the spec, and record fresh-context review verdicts for the four README-scoped requirements the rewrite invalidated (REQ-003.4.2, REQ-005.2.6, REQ-008.1.2, REQ-008.2.2). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XUFtq5xDd1rvWbW6Vqu5Ws
Collaborator
Author
|
For reviewers: this diff doubles as a walkthrough of the loop it documents. The README edit invalidated exactly the four verdicts that vouched for README content (REQ-003.4.2, REQ-005.2.6, REQ-008.1.2, REQ-008.2.2), CI held the merge until they were re-earned, and the four verdict JSONs in this diff are the fresh-context re-reviews that re-earned them — hash-bound to the new text. The first CI failure was real, too: REQ-011.8's tests caught the rewrite dropping README content the spec requires, and the fix was to satisfy the spec, not soften the test. |
…E-scoped verdicts Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XUFtq5xDd1rvWbW6Vqu5Ws
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.
Restructures the README for a first-time evaluator: the four-step loop up top, a real failing
check(captured from a live run), boundaries stated before the adoption instructions per REQ-008.1.2, the anti-cheat mechanics promoted with a real committed verdict inline, a positioning section vs. spec-generation tools and mutation testing, a short why-we-built-this, and the independent-runner pattern named in Risks (#32 tracks the executable example). The file-scoped-ID deep-dive moves todocs/file-scoped-ids.md; panopticon is linked as a live adoption example; the ~0.2s perf figure is replaced with a measured bound (<2s on a ~300k-line, 3,000-requirement corpus).Editing the README correctly invalidated four README-scoped review verdicts (REQ-003.4.2, REQ-005.2.6, REQ-008.1.2, REQ-008.2.2) — they need a fresh-context re-review per the usual flow.
🤖 Generated with Claude Code
https://claude.ai/code/session_01XUFtq5xDd1rvWbW6Vqu5Ws