fix: questions must show the content they refer to (overlay #879)#21
Merged
Conversation
…r to Generalizes the inline-approval rule (PR garrytan#1116) to every question in every module: an AskUserQuestion that asks the user to confirm, approve, rank, or choose among session-produced content (premises, findings, plans, scores) must render that content as direct assistant text first or restate it in the question itself. Internal reasoning and collapsed tool output do not count. Evidence: a /plan Discovery run asked 'Do you agree with all five premises?' with zero assistant text between the prior question and the ask — the premises were never shown (upstream issue garrytan#879 failure class, Phase 3). Pinned counts: 28 -> 29 overlays; EXPECTED_PARITY_CHECKS 4914 -> 5027 (2 x 55 modules + 3 regression checks). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
bun run gen:gstack2 output only — no hand edits. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
Bugbot is not enabled for your account, so this pull request was not reviewed. Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
A /plan Discovery session asked "Do you agree with all five premises?" with zero assistant text emitted since the previous question — the premises existed only in the agent's internal reasoning and were never shown to the user. This is upstream issue garrytan#879's failure class, previously patched only for the office-hours Phase-5 design doc (overlay garrytan#1116) and never generalized.
Fix
New judgment overlay garrytan#879 ("Self-contained questions") targeting
'*'— all 55 gstack2 modules: before any AskUserQuestion or prose decision brief about session-produced content (premises, findings, plans, approaches, scores, summaries), render that content as direct assistant text immediately before the question or restate it inside the question/options. Internal reasoning and collapsed tool output do not count as shown.scripts/gstack2/bug-fix-overlays.ts: overlay + executable regression (case 879)EXPECTED_PARITY_CHECKS4914 → 5027 (2 × 55 modules + 3 regression checks)pr-879.jsonfixture, provenance docsTesting
bun run test:gstack2— 26 test files green, full 5,027-check parity run passesbun test— free suite passes (exit 0)🤖 Generated with Claude Code
Summary by cubic
Make all user questions self‑contained by showing or restating the content they refer to before asking. Adds judgment overlay garrytan#879 across all modules to prevent invisible‑premise asks and resolves upstream issue garrytan#879.
evals/parity/regressions/pr-879.json.Written for commit ac65529. Summary will update on new commits.