v1.9.0 Pass 5 — finalize release (Apr 23 Anthropic post citation + release summary)#122
Merged
Merged
Conversation
…elease summary Closes the v1.9.0 cycle. Two small but load-bearing additions: K — Agent SDK credit-pool split. Already shipped in Pass 1 (TROUBLESHOOTING Models / API section) and Pass 2A (Cost-Conscious Composition callout). Pass 5 confirms coverage is complete — /coarse-review is a Claude Code plugin, not in our .claude/skills/, so no SKILL.md callout was possible; the TROUBLESHOOTING + guide entries fully document the 2026-06-15 cutover. L — Anthropic engineering-post citations: - Apr 8 2026 "Decoupling brain from hands" — already cited in Pass 3B (model-routing.md) and Pass 2A (Cost-Conscious Composition). - Apr 23 2026 "Quality reports update" — new callout added between the guide's "Mandatory Verification"/"Don't Skip Verification" warning and "Creating Your Own Domain Reviewer." Frames the post as evidence that model quality can regress and that the template's verification patterns (forked /verify-claims, passport.yaml, cross-artifact review, HIGH-WARN gate-refuse, /review-paper --variance N) all assume drift rather than treating any specific checkpoint as a stable baseline. Release finalisation: - CHANGELOG v1.9.0 heading: removed "(in progress)" marker. - Added "Inventory at release" line at the top of the v1.9.0 entry summarising the 30/14/24/6 → 36/16/26/6 delta and the 6 new skills / 2 new agents / 2 new rules. - Added "Release summary" table at the bottom of the v1.9.0 entry with one row per pass (PR # + net additions + what shipped). Verification: - check-surface-sync.sh: 26/26 pass, counts 36/16/26/6 - check-skill-integrity: pass - quarto render: clean - quality_score: 100/100 v1.9.0 is ready. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
There was a problem hiding this comment.
Pull request overview
Finalizes the v1.9.0 release documentation by adding an Anthropic Apr 23, 2026 engineering-post callout emphasizing model-quality regression risk and by updating the changelog to reflect the completed release state.
Changes:
- Added a verification-section callout in the workflow guide framing verification patterns as defenses against model-quality drift/regression.
- Updated rendered HTML outputs for the guide (including callout ID renumbering due to the insertion).
- Finalized the v1.9.0 CHANGELOG entry (removed “in progress”, added inventory-at-release + Pass 5 notes + release summary table).
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| guide/workflow-guide.qmd | Adds the new “model quality can regress” verification callout. |
| guide/workflow-guide.html | Regenerated rendered guide HTML to include the new callout and updated callout IDs. |
| docs/workflow-guide.html | Updates published docs HTML to match the regenerated guide output. |
| CHANGELOG.md | Finalizes v1.9.0 entry with inventory-at-release, Pass 5 closure notes, and a release summary table. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| The defences in this template all assume **model quality drifts**: | ||
|
|
||
| - **`/verify-claims`** with the Chain-of-Verification forked-verifier (the verifier cannot self-confirm even if the orchestrator's checkpoint regressed). | ||
| - **`/audit-reproducibility`** with `passport.yaml` (numeric claims are anchored to specific script output values, not to the model's recall of the values). |
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.
Summary
Closes the v1.9.0 cycle:
/coarse-reviewis a plugin, not in our.claude/skills/; no SKILL.md callout needed.Release finalisation:
Net release impact (v1.8.0 → v1.9.0)
New skills:
/humanize,/prompt,/prompt-only,/compress-session,/promote-memory,/stata-replication.New agents:
humanize-auditor,promote-memory-council.New rules:
model-routing.md,stata-code-conventions.md.New template:
passport-template.yaml. New reference:prompt-formatting-core.md.No breaking changes. Provenance: every addition traceable to the research-grounded plan at
quality_reports/plans/2026-05-20_v1.9.0-guide-refresh.md(local-only).Test plan
check-surface-sync.sh— 26/26check-skill-integrity— passquarto render— cleanquality_score— 100/100🤖 Generated with Claude Code