Hide branch scopes from the Explorer's scope picker

[mthines]

Summary

• Branch scopes (one-off PR/feature branches) churn constantly and were cluttering the Lore Explorer's scope picker at the top of the page — both the persistent chip strip and the "Browse all" searchable list.
• ScopeSelector now filters out type === 'branch' nodes from both surfaces. The fix lives in ScopeSelector itself rather than in buildScopeTree/flattenScopeTree (lib/scope-tree.ts), since those are also used by GroomingRuleBuilder's separate scope picker, which still needs branch scopes for pruning stale branch lessons.
• A branch scope that's the current selection (e.g. from a deep link) no longer gets a fallback chip in the strip either, consistent with branches being hidden from this picker entirely.

Test plan

• pnpm nx typecheck web — clean
• pnpm nx lint web — 0 errors (58 pre-existing warnings, unrelated to this change)
• npx vitest run --config vitest.storybook.config.ts src/components/lore/ScopeSelector — all 7 tests pass, including a new BranchScopesHidden interaction test asserting a branch scope appears in neither the strip nor Browse-all

https://claude.ai/code/session_01KzSdeCpiCbm19JM9NJCqkq

---

Generated by Claude Code

[dash0-dev]

✅ Reviewed your changes — no issues found.
_{Updated 2026-09-01 19:26 UTC}

Review details

^{Incremental review for commit 4f30775 (delta since ac24c9f).}

Gate	Status	Details
Description vs. code	✅	Description matches the diff exactly — scope, files, and rationale unchanged since the prior review.
Prior review feedback	✅	1 thread resolved since ac24c9f — no open review threads remain.
Documentation	✅	PR description and test plan clearly explain intent and behavior.
Self-review signals	✅	No debug logs, TODOs, or commented-out code in the delta.
Code review	✅	No new inline findings this run.

CI — All required checks pass (typecheck, lint, tests, Storybook, Vercel preview); several jobs report skipping (unaffected by this diff).

Run mode — incremental · 32 lines in delta

Memories — 331 indexed · 0 used

Quality — produced 0 → posted inline 0 (0 dropped, 0 deferred).

Integrations — not activated

Optimality (2.4c) — ran — delta only adds test coverage for an approach unchanged since the prior full review; no proposal met the materiality bar.

Standards (2.4d) — ran — delta is test-only (Storybook stories); no governing-doc violations found.

Skipped files — none

^{Reviewed by the pr-reviewer agent — open it to read how these gates and findings are produced.}

[github-actions]

The dashboard preview for this PR — redeployed on each push that changes the web app.

Name	Status	Preview	Deployment (this commit)	Updated (UTC)
lorekit	✅ Ready	Visit Preview	4f30775	2026-09-01 19:17

_{Preview always points at the PR's latest commit; Deployment is this exact commit. · Comment /web-preview to force a redeploy. · Workflow logs}

[mthines]

/web-preview

[github-actions]

❌ The dashboard preview for aa0361c failed to deploy (failure).

Open the run logs for the failure. Comment /web-preview to retry once it is fixed.

[dash0-dev]

Darkplane auto-approval is enabled for this repository (mode: Auto-approve).

An evaluation will run once this pull request's CI checks have completed — no action needed.

[dash0-dev]

Darkplane did not approve this pull request. It scored 69.21 of 100, and 75 was required.

• 🟡 Change footprint scored 45.
  The change touches ScopeSelector.tsx (+24/-12) with a new companion test story file (ScopeSelector.test.stories.tsx, +67/-0). The diff content itself could not be retrieved (tool returned no output), so the actual logic change is unverified; based on file scope and size, this looks like a small, localized UI filtering change to a single component plus added tests, but without seeing the diff I cannot confirm it avoids load-bearing logic.

The other 2 criteria

• 🟢 Blast radius scored 90.
  The change is confined to a single UI component ScopeSelector.tsx (a filter added to exclude type === 'branch' from topNodes/allScopes) plus its own test stories; no shared logic like flattenScopeTree or buildScopeTree is touched, and GroomingRuleBuilder's separate scope picker is explicitly left untouched. The diff is small (+24/-12 source, +67 test), purely additive filtering logic with no schema, API, or cross-component changes, making it easy to revert.
• 🟢 Evidence quality scored 72.
  All CI checks pass (typecheck, lint, Storybook interaction/visual tests, Vercel preview), and the diff added a new interaction test (BranchScopesHidden) plus a follow-up test (SelectedBranchGetsNoFallbackChip) addressing the automated reviewer's low-severity suggestion, showing the reviewer's feedback thread was resolved. Review coverage is from an automated bot only (no human reviewer signal), and detailed test assertions weren't independently inspected beyond commit history and CI status.

_{An AI evaluator produced these scores from the diff, the repository history, and the pull request text. The pull request text is author-controlled, so treat the scores as signals rather than proof. Reviewed for commit 4f30775.}