RomanNum3ral
/
gstack
mirror of https://github.com/garrytan/gstack.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
gstack/test
History
Garry Tan 5fe1814310
fix(plan-reviews): restore RECOMMENDATION + split Completeness by question type 
Opus 4.7 users reported /plan-ceo-review and /plan-eng-review stopped
emitting the RECOMMENDATION line and per-option Completeness: X/10
scores. E2E capture showed the real failure mode: on kind-differentiated
questions (mode selection, architectural A-vs-B, cherry-pick), Opus 4.7
either fabricated filler scores (10/10 on every option — conveys nothing)
or dropped the format entirely when the metric didn't fit.

Fix is at two layers:

1. scripts/resolvers/preamble/generate-ask-user-format.ts splits the old
   run-on step 3 into:
   - Step 3 "Recommend (ALWAYS)": RECOMMENDATION is required on every
     question, coverage- or kind-differentiated.
   - Step 4 "Score completeness (when meaningful)": emit Completeness: N/10
     only when options differ in coverage. When options differ in kind,
     skip the score and include a one-line explanatory note. Do not
     fabricate scores.

2. scripts/resolvers/preamble/generate-completeness-section.ts updates
   the Completeness Principle tail to match. Without this, the preamble
   contained two rules (one conditional, one unconditional) and the
   model hedged.

Template anchors reinforce the distinction where agent judgment is most
likely to drift:

- plan-ceo-review Section 0C-bis (approach menu) gets the
  coverage-differentiated anchor.
- plan-ceo-review Section 0F (mode selection) gets the kind-differentiated
  anchor.
- plan-eng-review CRITICAL RULE section gets the coverage-vs-kind rule
  for every per-issue AskUserQuestion raised during the review.

Regenerated SKILL.md for all T2 skills + golden fixtures refreshed. Every
skill using the T2 preamble now has the same conditional scoring rule.

Verified via new periodic-tier eval (test/skill-e2e-plan-format.test.ts):
all 4 cases fail on prior behavior, all 4 pass with this fix.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 		2026-04-22 01:11:05 -07:00
..
fixtures fix(plan-reviews): restore RECOMMENDATION + split Completeness by question type 2026-04-22 01:11:05 -07:00
helpers test: add AskUserQuestion format regression eval for plan reviews 2026-04-22 01:10:35 -07:00
analytics.test.ts feat: safety hook skills + skill usage telemetry (v0.7.1) (#189) 2026-03-18 23:57:59 -05:00
audit-compliance.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
benchmark-cli.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
benchmark-runner.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
builder-profile.test.ts feat: relationship closing — office-hours adapts to repeat users (v0.16.2.0) (#937) 2026-04-08 22:21:28 -10:00
codex-e2e.test.ts feat: worktree isolation for E2E tests + infrastructure elegance (v0.11.12.0) (#425) 2026-03-23 23:05:22 -07:00
codex-hardening.test.ts codex + Apple Silicon hardening wave (v0.18.4.0) (#1056) 2026-04-18 12:30:54 +08:00
context-save-hardening.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
diff-scope.test.ts feat: Review Army — parallel specialist reviewers for /review (v0.14.3.0) (#692) 2026-03-30 22:07:50 -06:00
explain-level-config.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
gemini-e2e.test.ts feat: Confusion Protocol, Hermes + GBrain hosts, brain-first resolver (v0.18.0.0) (#1005) 2026-04-16 10:41:38 -07:00
gen-skill-docs.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
global-discover.test.ts fix: close redundant PRs + friendly error on all design commands (v0.15.8.1) (#817) 2026-04-05 02:02:06 -07:00
gstack-developer-profile.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
gstack-question-log.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
gstack-question-preference.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
hook-scripts.test.ts feat: safety hook skills + skill usage telemetry (v0.7.1) (#189) 2026-03-18 23:57:59 -05:00
host-config.test.ts community wave: 6 PRs + hardening (v0.18.1.0) (#1028) 2026-04-17 00:45:13 -07:00
jargon-list.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
learnings-injection.test.ts fix: community security wave — 8 PRs, 4 contributors (v0.15.13.0) (#847) 2026-04-06 00:47:04 -07:00
learnings.test.ts feat: GStack Learns — per-project self-learning infrastructure (v0.13.4.0) (#622) 2026-03-29 17:02:01 -06:00
migration-checkpoint-ownership.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
openclaw-native-skills.test.ts community wave: 6 PRs + hardening (v0.18.1.0) (#1028) 2026-04-17 00:45:13 -07:00
plan-tune.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
readme-throughput.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
relink.test.ts fix: headed browser auto-shutdown + disconnect cleanup (v0.18.1.0) (#1025) 2026-04-16 15:39:44 -07:00
review-log.test.ts fix: community PRs + security hardening + E2E stability (v0.12.7.0) (#552) 2026-03-26 23:21:27 -06:00
setup-codesign.test.ts codex + Apple Silicon hardening wave (v0.18.4.0) (#1056) 2026-04-18 12:30:54 +08:00
ship-version-sync.test.ts fix(ship): detect + repair VERSION/package.json drift in Step 12 (v1.1.1.0) (#1063) 2026-04-18 23:58:59 +08:00
skill-collision-sentinel.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
skill-e2e-autoplan-dual-voice.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
skill-e2e-benchmark-providers.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
skill-e2e-bws.test.ts fix: cookie picker auth token leak (v0.15.17.0) (#904) 2026-04-08 10:10:13 -07:00
skill-e2e-context-skills.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
skill-e2e-cso.test.ts feat: /cso v2 — infrastructure-first security audit (v0.11.6.0) (#384) 2026-03-23 06:57:22 -07:00
skill-e2e-deploy.test.ts feat: /land-and-deploy first-run dry run + staging-first + trust ladder (v0.12.2.0) (#518) 2026-03-26 11:08:31 -07:00
skill-e2e-design.test.ts feat: CI evals on Ubicloud — 12 parallel runners + Docker image (v0.11.10.0) (#360) 2026-03-23 10:17:33 -07:00
skill-e2e-learnings.test.ts feat: recursive self-improvement — operational learning + full skill wiring (v0.13.8.0) (#647) 2026-03-31 23:08:22 -06:00
skill-e2e-office-hours.test.ts feat: mode-posture energy fix for /plan-ceo-review and /office-hours (v1.1.2.0) (#1065) 2026-04-19 05:44:39 +08:00
skill-e2e-plan-format.test.ts test: add AskUserQuestion format regression eval for plan reviews 2026-04-22 01:10:35 -07:00
skill-e2e-plan-tune.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
skill-e2e-plan.test.ts feat: mode-posture energy fix for /plan-ceo-review and /office-hours (v1.1.2.0) (#1065) 2026-04-19 05:44:39 +08:00
skill-e2e-qa-bugs.test.ts feat: CI evals on Ubicloud — 12 parallel runners + Docker image (v0.11.10.0) (#360) 2026-03-23 10:17:33 -07:00
skill-e2e-qa-workflow.test.ts feat: CI evals on Ubicloud — 12 parallel runners + Docker image (v0.11.10.0) (#360) 2026-03-23 10:17:33 -07:00
skill-e2e-review-army.test.ts feat: Review Army — parallel specialist reviewers for /review (v0.14.3.0) (#692) 2026-03-30 22:07:50 -06:00
skill-e2e-review.test.ts feat: Confusion Protocol, Hermes + GBrain hosts, brain-first resolver (v0.18.0.0) (#1005) 2026-04-16 10:41:38 -07:00
skill-e2e-session-intelligence.test.ts fix(checkpoint): rename /checkpoint → /context-save + /context-restore (v1.0.1.0) (#1064) 2026-04-19 08:38:19 +08:00
skill-e2e-sidebar.test.ts feat: declarative multi-host platform + OpenCode, Slate, Cursor, OpenClaw (v0.15.5.0) (#793) 2026-04-04 15:32:20 -07:00
skill-e2e-workflow.test.ts refactor: extract TabSession for per-tab state isolation (v0.15.16.0) (#873) 2026-04-07 00:23:36 -07:00
skill-e2e.test.ts feat: recursive self-improvement — operational learning + full skill wiring (v0.13.8.0) (#647) 2026-03-31 23:08:22 -06:00
skill-llm-eval.test.ts feat: voice directive for all skills (v0.12.3.0) (#520) 2026-03-26 17:31:53 -06:00
skill-parser.test.ts feat: SKILL.md template system, 3-tier testing, DX tools (v0.3.3) (#41) 2026-03-13 21:08:12 -07:00
skill-routing-e2e.test.ts feat: Confusion Protocol, Hermes + GBrain hosts, brain-first resolver (v0.18.0.0) (#1005) 2026-04-16 10:41:38 -07:00
skill-validation.test.ts feat(v1.4.0.0): /make-pdf — markdown to publication-quality PDFs (#1086) 2026-04-20 13:20:30 +08:00
taste-engine.test.ts feat(v1.3.0.0): open agents learnings + cross-model benchmark skill (#1040) 2026-04-19 17:50:31 +08:00
team-mode.test.ts feat: Confusion Protocol, Hermes + GBrain hosts, brain-first resolver (v0.18.0.0) (#1005) 2026-04-16 10:41:38 -07:00
telemetry.test.ts feat: community wave — 7 fixes, relink, sidebar Write, discoverability (v0.13.5.0) (#641) 2026-03-29 21:43:36 -06:00
timeline.test.ts feat: Session Intelligence Layer — /checkpoint + /health + context recovery (v0.15.0.0) (#733) 2026-04-01 00:50:42 -06:00
touchfiles.test.ts test: add AskUserQuestion format regression eval for plan reviews 2026-04-22 01:10:35 -07:00
uninstall.test.ts feat: community PRs — faster install, skill namespacing, uninstall, Codex fallback, Windows fix, Python patterns (v0.12.9.0) (#561) 2026-03-27 00:44:37 -06:00
upgrade-migration-v1.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
v0-dormancy.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00
worktree.test.ts feat: content security — 4-layer prompt injection defense for pair-agent (#815) 2026-04-06 14:41:06 -07:00
writing-style-resolver.test.ts feat: gstack v1 — simpler prompts + real LOC receipts (v1.0.0.0) (#1039) 2026-04-18 15:05:42 +08:00