v1.100.0: Reliable default — Opus 4.6[1m] + GPT-5.5 + Fable - #712
Merged
Merged
Conversation
Rename every "Setup A is the default" / "recommended default, trial" occurrence to either "Reliable — recommended default" or "bleeding edge". Files: CLAUDE_CODE_SDLC_WIZARD.md, skills/setup/SKILL.md, skills/update/SKILL.md, tests/test-doc-consistency.sh, AI_SETUP_LANES.md Codex review: round 2 CERTIFIED SOUND 98%, round 3 recheck CERTIFIED SOUND 96% (tree c27be39). Zero findings across both rounds.
Codex full-release review (GPT-5.5 xhigh) found README.md:214 still said "Four Setup Lanes" and AI_SETUP_LANES.md:223 said "all four lanes" after the Reliable row was added to the table. Both now say five.
- skills/sdlc/SKILL.md: Recommended Model now leads with Opus 4.6[1m] max (Reliable default), Opus 5 high listed as bleeding edge - Plugin manifests bumped 1.99.2 → 1.100.0 - Cowork copy synced - Test anchor updated: Opus 5 → Opus 4.6 with max effort check
The helper matched fixed-string `max` anywhere on the line, which also hit "not blanket `max`" later in the same sentence. Mutating Opus 4.6 max→high did not fail the test. Replace with direct grep for the exact pattern. TDD RED proven: max→high mutation fails, revert passes. 137/0.
The launcher hardcoded -m gpt-5.5. Running Sol or any other model required copying the script and patching paths — which broke three times in a row. Now: REVIEW_MODEL=gpt-5.6-sol scripts/run-review-leg.sh OUT PROMPT Default unchanged (gpt-5.5). TDD RED/GREEN proven. 36/0 tests.
Sol (GPT-5.6 xhigh) found: - 3 stale 1.99.2 references in shipped files (wizard doc, update skill) - test-escalation-ladder.sh grep broke on env-var launcher pattern - default-model test could inherit REVIEW_MODEL from caller env Disputed: auto-mode [N] default and gpt-5.6-sol reviewer role are by design, not Reliable lane contradictions. 36/0 + 6/0 + 137/0.
CI caught SDLC.md version marker still at 1.99.2. ROADMAP.md had the same in its last-release line and dogfood tracker table.
The test banned any "95%" in the setup skill. The Reliable lane's user message says "escalation brain at <95% confidence" — a brain ladder description, not the setup wizard's own state model. Narrowed the check to confidence-THRESHOLD patterns only.
Owner
Author
|
CROSS-MODEL-CLEARANCE {"reviewer": "gpt-5.5", "verdict": "YES", "confidence": 97, "sha": "5ee0a2340cbfa7abe11780ab3578b0137c40738e"}Codex (GPT-5.5 xhigh): CERTIFIED SOUND across rounds 2-5 (96-98%). Zero blocking findings. |
Owner
Author
|
CROSS-MODEL-CLEARANCE {"reviewer": "fable", "verdict": "YES", "confidence": 98, "sha": "5ee0a2340cbfa7abe11780ab3578b0137c40738e"}Fable (advisor): history clean, no Opus-5 code contamination. Ruled release strategy sound. |
Owner
Author
|
USER-APPROVED MERGE OVERRIDE Reason: v1.100.0 Reliable default release — user gave 100% clearance Waived (would otherwise have blocked):
Still verified, not waivable: CI Head: Posted by |
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.
Summary
[r] Reliableoption to setup skill Step 9.5 and update skill Fable-as-driver promptFiles changed
package.json— v1.99.2 → v1.100.0README.md— version badge, Reliable lane descriptionCHANGELOG.md— v1.100.0 entryAI_SETUP_LANES.md— Setup A label fixCLAUDE_CODE_SDLC_WIZARD.md— effort table, driver/advisor table, opt-in section, autocompact, latest tier, escape hatchskills/setup/SKILL.md— Step 9.5[r]Reliable optionskills/update/SKILL.md— Fable-as-driver prompt, effort check orderingtests/test-doc-consistency.sh— autocompact guard grep updatedCross-model review
Test plan
tests/test-doc-consistency.sh— 137/0tests/test-cli.sh— 97/0 (run by codex during review)validateworkflow