⚡ Bolt: [성능 개선] skip-link 애니메이션 최적화 - #148

Pull request overview

OpenCode reviewed the current-head bounded evidence and requested changes before merge.

Findings

1. P2 styles.css:821 - Missing regression test for the changed .skip-link animation surface

• Problem: The PR changes the .skip-link show/hide animation from the layout property top to transform (top: 0; transform: translateY(-100%); with transition: transform 0.2s ease-in-out) but adds no regression test, even though tests/test_styles.py is the repository's dedicated suite for performance-sensitive site CSS and the 2026-08-05 .jules/bolt.md entry explicitly requires non-vacuous tests for exactly this class of change.
• Root cause: The PR contains only a CSS edit plus CHANGELOG and learning-log entries; no test surface was touched, leaving the changed rule outside the coverage of the only styles.css consumer (tests/test_styles.py, per CodeGraph blast radius).
• Fix: Add a static regression test to tests/test_styles.py using the existing _rule() helper that asserts .skip-link declares transform: translateY(-100%); and transition: transform and never animates top (assert 'transition: top' not in rule).
• Regression test: Add test_skip_link_animates_transform_not_top to tests/test_styles.py and run: python3 -m pytest tests/test_styles.py -q
• Suggested diff: posted in this finding's inline review thread.

Summary

PR #148 changes styles.css .skip-link to animate transform: translateY(-100%) -> translateY(0) instead of the layout property top (top: 0 plus translateY(-100%) keeps the link fully off-screen at any height; :focus-visible reveals it; the existing global prefers-reduced-motion rule cuts transition-duration to 0.01ms so reduced-motion users get an instant reveal with no forced animation). CHANGELOG.md:4 and .jules/bolt.md:20-22 match the styles.css head bytes. Changed files inspected: styles.css (hunk @@ -809,37 +809,39 @@), CHANGELOG.md, .jules/bolt.md. Blocking P2: no regression test was added although tests/test_styles.py is the repo's dedicated suite for performance-sensitive site CSS and the 2026-08-05 .jules/bolt.md entry explicitly requires non-vacuous tests for this class of change; CodeGraph blast radius shows STYLES (tests/test_styles.py:9) as the only styles.css consumer and that file has no .skip-link/transform/transition assertion, so a reintroduced transition: top would ship undetected to the live GitHub Pages site. Approval sufficiency: partial - CSS correctness confirmed by source trace, but verification coverage for the changed surface is missing. Verification posture: trusted source trace, current-head diff, CodeGraph blast radius; no OPENCODE_EXECUTION_RECEIPT available for this PR. Linter/static: only hadolint Dockerfile configured; no applicable lint for styles.css. TDD/regression: missing - no test added for the changed CSS rule. Coverage: not applicable - Coverage execution evidence PASS because no supported changed source files or package manifests were found. Docstring coverage: not applicable - same Coverage execution evidence basis. DAG: head-state flowchart in the review body maps styles.css .skip-link to focus-visible reveal, reduced-motion 0.01ms fallback, and the test_styles.py coverage gap. PoC/execution: none - no execution receipt exists; render behavior judged by source trace only. DDD/domain: static personal site; no domain model affected. CDD/context: no context boundaries affected. Similar issues: prior PR #141 and .jules/bolt.md 2026-08-05 established the repo convention of non-vacuous CSS regression tests, which this PR does not follow. Claim/concept check: CHANGELOG and bolt.md claims match current-head styles.css; only the absolute '0ms' main-thread claim in .jules/bolt.md:22 is unmeasured (P3 note). Standards search: compositor-only transform animation for off-screen skip-link reveal is the documented Web Platform best practice; reduced-motion fallback present. Compatibility/convention: no new identifiers, selectors, or reserved words introduced; .skip-link name unchanged. Breaking-change/backcompat: none - static CSS only. Implementation completeness: CSS behavior is complete; the gap is regression-test coverage, not a placeholder body. Performance: transform-based animation avoids per-frame layout reflow; the absolute '0ms' claim is unmeasured. Developer experience: DX surface is the static-site CSS edit plus pytest regression workflow (tests/test_styles.py); judged against repo convention. User experience: UX surface is the keyboard skip-link focus reveal; verified by source trace only, no DOM receipt. Visual/DOM: no Playwright/browser execution receipt in bounded evidence; reviewed the non-web static CSS/docs/test interaction surface instead. Accessibility/i18n: skip link remains keyboard-reachable via :focus-visible; prefers-reduced-motion honored; Korean copy unchanged. Supply-chain/license: no dependency changes. Packaging: no manifests changed; unpackaged_source_surfaces flags pre-existing node (components/krds-gallery.js, i18n.js) and python (tests/) surfaces unchanged by this PR. Security/privacy: no security surface changed; no sequential-id or identifier-exposure surface introduced.

Adversarial validation

{"status":"failed","probes":[{"path":"styles.css","line":821,"hypothesis":"A regression reintroducing top-based animation of .skip-link (e.g. transition: top 0.2s ease-in-out or top: -40px) would ship undetected because no repository test covers the changed rule.","attack_or_counterexample":"Revert the .skip-link hunk at styles.css:821-836 to the pre-PR top-animation form and run the existing regression suite.","evidence":"Trusted source trace at styles.css:821 observed that the current-head Changed files section lists exactly .jules/bolt.md, CHANGELOG.md, styles.css with no test file; CodeGraph blast radius names STYLES (tests/test_styles.py:9) as the only styles.css consumer; verbatim tests/test_styles.py (lines 1-104) contains no .skip-link, transform, or transition assertion, so the suite passes (exit 0) unchanged against the reverted counterexample; source-line-sha256=ee22eb69570e01b0f0ba7c37679dfeedf94ad21ca00098361988d814cb0d69af","outcome":"confirmed"},{"path":"styles.css","line":835,"hypothesis":"Users with prefers-reduced-motion: reduce lose the skip link or are forced through animated motion because the transition is suppressed.","attack_or_counterexample":"OS-level reduced-motion setting while a keyboard user tabs to the skip link.","evidence":"Trusted source trace at styles.css:835 observed the focused changed hunks retain the global @media (prefers-reduced-motion: reduce) rule setting transition-duration: 0.01ms !important on *, *::before, *::after, so the :focus-visible rule still applies transform: translateY(0) instantly and the link is revealed without forced smooth animation - the counterexample (lost or forced-motion skip link) is rejected; source-line-sha256=7a82d67b666849def665c0d1c8323fef29f6709e7215bdb448eca119f81c9cc1","outcome":"falsified"},{"path":"CHANGELOG.md","line":4,"hypothesis":"The CHANGELOG and .jules/bolt.md entries describe behavior the styles.css head does not implement (doc-to-code mismatch).","attack_or_counterexample":"Compare the CHANGELOG.md:4 claim (top -> transform: translateY()) and the .jules/bolt.md:20-22 entry against the styles.css diff.","evidence":"Trusted source trace at CHANGELOG.md:4 observed the focused changed hunks of styles.css declare top: 0; transform: translateY(-100%); in .skip-link, transform: translateY(0) in .skip-link:focus-visible, and transition: transform 0.2s ease-in-out, matching the CHANGELOG entry and .jules/bolt.md:20-22 verbatim - the doc-to-code mismatch hypothesis is rejected; source-line-sha256=acae470d708fb848d4890034f71992e9a11d0255f82f484410162851e22f1fcb","outcome":"falsified"}],"residual_risk":"Confirmed risk is regression-test coverage: the .skip-link rule can silently regress to top-based animation until deployed to the live site. Rendering of the transform transition was not executed (no browser/DOM execution receipt in bounded evidence); correctness is established by source trace and the repo's static-test convention only. mergeStateStatus BLOCKED is branch policy, not a merge conflict (mergeable=true)."}

• Result: REQUEST_CHANGES

• Reason: The .skip-link animation change in styles.css ships without the regression test the repository requires for performance-sensitive CSS; the changed rule is uncovered by tests/test_styles.py, the only consumer of styles.css per CodeGraph blast radius.

• Head SHA: 36e2b7f7d4772529bfc73ab593b1111a7ab137ef

• Workflow run: 31320130184

• Workflow attempt: 1

Changed-File Evidence Map

flowchart LR
  PR["PR changed files"] --> Evidence["OpenCode bounded evidence"]
  Evidence --> S1["Changed file (3 files)"]
  S1 --> I1["repository behavior"]
  I1 --> R1["Review risk: Changed file (3 files)"]
  R1 --> V1["required checks"]

Actionable comments posted: 2

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In @.jules/bolt.md:
- Around line 20-22: Revise the `.jules/bolt.md` entry to remove guarantees
about GPU acceleration and 0ms main-thread cost, distinguishing observed absence
of Layout events from general browser behavior and describing transform as
reducing layout recalculation. Also update `CHANGELOG.md` at line 4 to replace
“GPU acceleration rendering” with qualified wording that the transform
transition can reduce layout recalculation.

In `@tests/test_styles.py`:
- Around line 107-113: Strengthen test_skip_link_animates_transform_not_top by
extracting the transition declaration body before asserting that top is absent,
so combined transform/top transitions fail. Also inspect the skip-link focus
state and assert it uses transform with translateY(0).

Pull request overview

OpenCode reviewed the current-head bounded evidence and found no blocking issues.

Findings

No blocking findings.

Summary

Approval sufficiency: current-head evidence affirmatively supports the PR intent — Coverage execution evidence Result: PASS (supported repository test suites passed), Failed GitHub Check evidence reports no completed failed checks, Other unresolved review thread evidence reports none; changed files inspected: styles.css and tests/test_styles.py (plus docs .jules/bolt.md, CHANGELOG.md). Verification posture: static-site CSS change verified by repo-native Python static tests in tests/test_styles.py (test_skip_link_animates_transform_not_top added at head f6cd041); Coverage execution evidence Result: PASS confirms the supported suite passed; no OPENCODE_EXECUTION_RECEIPT browser lines exist, so runtime visual behavior is a stated limitation, not claimed. Linter/static: only hadolint Dockerfile and trivy fs . contracts are configured; no lint/security failure evidence at head. TDD/regression: the new test is non-vacuous — assertions require transform: translateY(-100%) in the base rule, transform in the transition declaration, no 'top' in the transition, and transform: translateY(0) in the focus-visible rule, all of which fail against the base top-based rules; this directly resolves the prior opencode-agent CHANGES_REQUESTED thread (SHA 36e2b7f) which is outdated at current head. Coverage: Coverage execution evidence Result: PASS, supported repository test suites passed. Docstring coverage: configured repository docstring gates passed or advisory per Coverage execution evidence; the new test carries a docstring. DAG: flowchart of the .skip-link focus flow (base rule -> transform: translateY(-100%) hidden state -> :focus-visible -> translateY(0), plus the prefers-reduced-motion 0.01ms override and the regression-test path), reflecting the base-to-head changed flow per CodeGraph and focused hunks. PoC/execution: no trusted execution receipts beyond Coverage Result: PASS; only source traces, diff hunks, and the coverage decision are cited. DDD/domain: no domain-logic change; purely presentational CSS. CDD/context: no context/API change. Similar issues: prior reviewer request for skip-link regression coverage is satisfied at tests/test_styles.py:107-118. Claim/concept check: PR, CHANGELOG, and .jules/bolt.md claims are explicitly hedged (browser-measured effects, no GPU/0ms guarantees), consistent with the actual CSS change and with the repo's documented image-hint measurement-limits convention. Standards search: transform-based offscreen skip-link and the global prefers-reduced-motion override (transition-duration: 0.01ms !important) are standard, source-consistent patterns; no contradicting evidence. Compatibility/convention: no new externally meaningful identifiers, DB objects, or API fields; only pre-existing .skip-link selectors and a descriptive multi-word test function name; no reserved-word or naming concerns. Breaking-change/backcompat: no public contract change; CSS and test files only. Performance: animating transform instead of top avoids per-frame layout work during the transition, matching the PR intent; measured effect correctly deferred to per-browser measurement in docs. Developer experience: static tests keep the no-build pytest surface; unpackaged_source_surfaces notes tests/*.py have no package manifest, but the supported suite passes per Coverage evidence and this PR adds no tooling. User experience: skip-link remains fully hidden at rest (translateY(-100%) above the initial containing block) and appears at the top-left on focus-visible; reduced-motion users receive 0.01ms transitions via the existing media query. Visual/DOM: non-web interaction surface reviewed — static CSS source, focused diff hunk styles.css @@ -809,37 +809,39 @@, and Python rule-parsing tests; no browser/DOM receipts exist so no runtime DOM observation is claimed. Accessibility/i18n: focus-visible behavior preserved and contract-tested; reduced-motion fallback present; i18n unaffected. Supply-chain/license: no dependency changes. Packaging: no package/build changes; CSS/docs/test-only PR. Security/privacy: no auth, data, input, or identifier surface changed; no enumeration or injection concerns.

Approval sufficiency: bounded evidence supplied affirmative approval evidence for changed files, coverage/docstring posture, risk surfaces, and current-head verification; approval is not based merely on the absence of known blockers.
Verification posture: CodeGraph evidence was initialized and bounded current-head evidence reviewed for changed-file evidence including .jules/bolt.md, CHANGELOG.md, styles.css, tests/test_styles.py.
Linter/static: workflow/static review evidence is bounded by the current-head GitHub Checks gate and changed-file evidence.
TDD/regression: coverage execution evidence and focused changed hunks were reviewed from bounded-review-evidence.md.
Coverage: coverage execution evidence reports supported repository test suites passed.
Docstring coverage: coverage execution evidence reports configured repository docstring gates passed or docstring coverage was advisory.
DAG: CodeGraph/source-backed behavior map connects .jules/bolt.md to the affected review, runtime, or workflow path and required checks.
PoC/execution: coverage-evidence job executed on the current head and reported PASS.
DDD/domain: workflow and repository-governance invariants were reviewed against changed files in bounded evidence.
CDD/context: CodeGraph evidence, changed-file history, and focused hunks were reviewed from bounded-review-evidence.md.
Similar issues: changed-file history evidence was reviewed for comparable local precedents.
Claim/concept check: bounded evidence, repository source, current-head workflow evidence, and, where numeric, scientific, statistical, or literature-backed claims are affected, original-paper/formula evidence and parameter-recovery expectations were used for claims.
Standards search: standards and external-source claims require trusted bounded source evidence prepared outside the isolated model process; no evidence-backed standards blocker is present in bounded evidence.
Compatibility/convention: changed workflow/script conventions, object naming, and reserved-word safety for schema/API/config/code surfaces were checked in bounded evidence.
Breaking-change/backcompat: deployment evidence and changed-file history were checked for backward-compatibility risk.
Performance: changed surfaces were checked for performance risk in bounded evidence.
Developer experience: changed automation, review, test, setup, and maintenance surfaces were checked for helpful or obstructive DX impact in bounded evidence.
User experience: connected user, operator, API, CLI, documentation, review-comment, status-check, rendering, and workflow-reader behavior was checked for contradictions against code, docs, and tests in bounded evidence.
Visual/DOM: deterministic repair does not infer browser runtime execution; source-backed DOM/UI evidence and trusted workflow receipts were reviewed when present, and non-web surfaces used API/CLI/log/docs/workflow evidence instead.
Accessibility/i18n: accessibility, localization, and human-readable text surfaces were checked where UI, CLI, API message, docs, logs, or review text changed.
Supply-chain/license: dependency, package, model, container, and external-tool changes were checked in bounded evidence.
Packaging: package, build, test, lint, and security contracts were checked in bounded evidence.
Security/privacy: workflow-token, review-gate, and repository-automation security/privacy boundaries were checked in bounded evidence.

Adversarial validation

{"status":"passed","probes":[{"path":"styles.css","line":821,"hypothesis":"The resting skip-link becomes visible or overlaps top-of-page content after replacing top:-40px with top:0 plus transform:translateY(-100%), regressing the hidden-at-rest state","attack_or_counterexample":"Page loaded at scroll position 0, plus a revert to the base top-based animation rules","evidence":"Trusted focused diff at styles.css:821 (hunk @@ -809,37 +809,39 @@) shows the .skip-link base rule with position:absolute; top:0; left:0; transform:translateY(-100%) and transition:transform 0.2s ease-in-out while the :focus-visible rule restores transform:translateY(0); because the element is absolutely positioned and removed from flow, translateY(-100%) offsets it fully above the initial containing block so the resting state stays hidden with no overlap or layout shift, and the same head-file source is what the passing suite parses; Coverage execution evidence Result: PASS confirms the supported suite passed at head f6cd0417; source-line-sha256=ee22eb69570e01b0f0ba7c37679dfeedf94ad21ca00098361988d814cb0d69af","outcome":"falsified"},{"path":"tests/test_styles.py","line":118,"hypothesis":"The new regression test passes vacuously and would not catch a revert to the top-based transition","attack_or_counterexample":"Revert styles.css to base: .skip-link{top:-40px; transition:top 0.2s ease-in-out} and .skip-link:focus-visible{top:0} with no transform declarations","evidence":"Trusted source trace at tests/test_styles.py:118: the assertion assert \"transform: translateY(0);\" in focus_rule fails against the base focus-visible body top:0, the base-rule assertions require transform: translateY(-100%) and forbid 'top' in the transition, and _rule() itself asserts the selector exists, so the base version cannot pass the suite; Coverage execution evidence Result: PASS confirms the head suite including tests/test_styles.py executed and passed at head f6cd0417 (observed pass), so the assertions held against the current-head rule bodies; source-line-sha256=2c085de6eb5914cd7498bfa790f5bc2de1290aafdf129f859f1126b2850020c7","outcome":"falsified"}],"residual_risk":"CSS/docs/test-only change for a static GitHub Pages site. No browser-execution receipts exist in the bounded evidence, so real focus-traversal rendering, per-browser compositing behavior, and the transition's measured effect remain unverified at runtime; the repository's static Python suite (Coverage Result: PASS) and source-level contract tests are the only execution evidence. No security, data, or API surface is affected."}

• Result: APPROVE
• Reason: CSS transition change from top to transform with a matching non-vacuous regression test; Coverage execution evidence PASS, no active failed checks, no unresolved review threads at current head.
• Head SHA: f6cd0417b44d101e5c085d9c8f59d88a22c1bfe6
• Workflow run: 31352046950
• Workflow attempt: 1