## What
Two fixes from adversarial testing of the DE parallel router (10 hostile comps, routed vs screenshot-baseline PSNR). The router itself held — both bugs are in general drawElement fast capture, and one slipped past self-verify.
### 1. Compile gate: ancestor background-image (`producer`)
`drawElementService`'s per-frame ancestor fill replicates what lies behind the captured subtree by walking up the DOM for the nearest non-transparent **`backgroundColor`**. A background-**image** (`linear-gradient`, `url()`) on `body`/`html`/a wrapper reads as transparent in that scan, so a deeper ancestor's solid color paints instead wherever the subtree leaves pixels uncovered.
Measured repro: body `linear-gradient` + html solid color + an element shrinking late in the comp → DE paints the html purple instead of the body gradient. 30.9 dB min frame vs baseline, visually unmistakable. Identical damage single-worker and parallel — general DE bug, in every wild DE render matching this (very common) authoring pattern.
Fix: `detectAncestorBackgroundImage()` in the compiler (DOM-aware — inline styles on the root's ancestor chain + `<style>` rules resolved via `querySelectorAll`, so class-selected wrappers are covered; backgrounds *inside* the root are deliberately not matched). New compile gate `ancestor_background_image`, same shape as the 3D/mix-blend gates, bypass `HF_FAST_CAPTURE_ANCESTOR_BG=true`.
### 2. Self-verify tail sample (`engine`)
The verify grid sampled at `(i+1)/(k+1)` → [20/40/60/80]% of the timeline. The damage above starts at ~79% and peaks after the last sample — verification **passed** on output that bottomed at 30.9 dB (threshold 32 dB would have caught it, it just never looked there).
Fix: `computeDeVerifySampleFractions()` — first k−1 samples evenly spaced, last pinned at 95%. Default k=4 grid becomes [25/50/75/95]%. Kills the whole late-onset damage class, not just this repro.
## Validation
- Repro comp (body gradient + shrink reveal): now gates → baseline route, 54.7 dB avg vs ground truth (was 41.6 avg / 30.9 min with the purple surround)
- Control comp (nested stacked fades, routed): still routes, verify grid `[90, 180, 270, 342] of 360`, passes, 60.6 dB avg — unchanged
- Full adversarial matrix context: 6/10 comps routed clean (49–68 dB min), blend/3D gated correctly, animated-canvas damage caught by verify at 16.5 dB with clean revert, video comps route legitimately (frames pre-extracted)
- Tests: 7 new detection cases (`htmlCompiler.test.ts`), 5 new grid cases (`frameCapture-verifySampleFractions.test.ts`); `compileStage.test.ts` + `frameCapture.test.ts` suites green
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Selecting a rotated (cropped) element appeared to straighten it: the crop
dim and dashed window were drawn on the axis-aligned bounding box, so the
bright window was a straight rectangle and the element's rotated corners
were masked to near-black — while the DOM transform was untouched.
clip-path applies in the element's LOCAL frame, before its transform, so
the crop visualization now renders inside a container rotated with the
element: readElementCropFrame decomposes the computed 2D matrix into
angle + per-axis scale (element scale finally factored into the px
mapping too) and 3D/unparseable transforms keep the axis-aligned
presentation. Pointer deltas rotate into the element frame before the
inset resolvers, so edge/pan drags track the rotated handles correctly.
Review follow-ups (both reviewers, all findings):
- resize captures scope to the resize group: convert-to-keyframes
resolvedFromValues and the whole-offset backfill pass the group filter,
so an opacity-touching intro tween can't ride into a converted scale
tween (the rotation fix's contract, now uniform across intercepts)
- commitStaticSet resolves every group's target set BEFORE committing and
coalesces groups landing on the same legacy mixed set into one commit —
the second commit can no longer chase a stale group-derived id
- installAuthoredOpacityCapture also stamps an element the moment it GAINS
data-color-grading at runtime (attributeFilter), not just at insertion
- both writer twins now share the same emitted-set dedupe shape
- applySoftReload's positional tail becomes a SoftReloadOptions object
- readAllAnimatedProperties builds the group-filtered key set immutably
instead of deleting from the set mid-iteration
- applyAuthoredInlineOpacity documents the priority-lossy round-trip
- the marquee hit-test reads activeCompositionPathRef like its neighbors
New tests: resize intercept (scale route + group filter + non-uniform
longhands), after-write-HTML / stamp / empty-stamp opacity restore, the
no-op-commit-with-missed-instant-patch soft-reload contract, and the
runtime-gained-grading stamp.
The deselect restore read a ref recomputed from RENDER state — on a direct
A→B selection switch, state re-syncs to B before A's effect cleanup runs,
so after a committed crop gesture A was restored with B's crop string (or
lost its crop when B had none). The committed value is now written at
gesture-commit time (tri-state: none committed / crop removal / the exact
committed string), so cleanup never touches render state. Adds component
tests for lift/restore ordering, including the direct A→B switch and the
uneditable-clip stand-down.
gsap_from_opacity_noop matched 'opacity: 0' as a prefix — an authored
'opacity: 0.98' triggered the rule. One boundary-anchored predicate now
owns the exactly-zero test for both block and inline declarations,
including a block's last declaration without a trailing semicolon.
domEditingDom now imports the grading contract attribute from core
instead of re-declaring the literal.
A parsed timeline set carries immediateRender in extras; the recast
statement builder ALSO pushed the flag unconditionally, so every
split/re-add of a set wrote 'immediateRender: true, immediateRender:
true' into the file, doubling on each pass. Both writer twins (recast +
acorn) now emit each vars key exactly once, properties winning over
extras, and reconcileEditableProps skips newProps keys it already
preserved.
commitStaticSet merged every property into the FIRST set found for the
selector: a panel W edit on an element whose only set was positional
produced tl.set("#el",{x,y,width}) — a mixed-group set the split
machinery exists to prevent — labeled "Set 3D transform" in undo.
Commits now batch per property group into a set that owns that group
(exact group match, then a mixed set already carrying the group, then a
fresh off-timeline gsap.set), with undo labels derived from the group
(Move layer / Resize layer / Rotate layer / Set 3D transform).
The always-on crop lifted EVERY selected element's clip-path and restored
only what it could parse as a px inset — selecting an element with a
circle/polygon/percentage clip visually un-clipped it, and deselecting
deleted the authored clip outright.
readElementCropInsets is now tri-state (zeros = no clip, null = a clip
the tool can't represent): uneditable clips get no lift and no handles,
and the lift restores the pre-lift inline value verbatim unless a crop
gesture actually committed.
- resize on a scale-driven element commits per-axis scale (scaleX/scaleY
for non-uniform drags) with keyframe normalization to the longhands, and
clears the width/height draft so size can't double-apply; the intercept
moves to gsapResizeIntercept.ts
- the drop frame applies the corrected position synchronously in the same
microtask chain as the soft reload (no network-window jump), and the
draft pins the anchor through accumulated moves on scaled elements
- gesture size/position math divides by the element's own content scale
- convert-to-keyframes resolves current values through the property-group
filter for ALL capture passes (opacity/rotationX from unrelated tweens
no longer leak into a rotation commit), and a grading-hidden source's
opacity is read from its canvas, not the inline hide
- canvas pointer-down confirms the hover target with a synchronous
hit-test before starting a marquee (stale-hover race lost selections)
Editing commits made elements vanish or dim permanently: invalidating the
whole timeline (or re-running the composition script on soft reload) made
GSAP re-capture tween bounds while runtime transients were live — the
grading hide's opacity 0, or a mid-flight tween value — so from()/to()
bounds got poisoned and the element rendered invisible from then on.
- patch only the edited tween in place, never timeline.invalidate()
- soft reload restores every animated element's authored inline opacity
(after-write HTML first, parse-time stamp as fallback) before the script
re-runs and re-captures
- a paired x/y commit whose second half is a no-op (changed=false) still
applies its instant patch, so panel edits reflect without deselecting
The color-grading engine hides its source element with inline
'opacity: 0 !important', so any code that later reads or re-captures the
element's opacity sees the hide instead of the authored value. Stamp the
authored inline opacity on every [data-color-grading] element at document
parse time (MutationObserver installed at runtime-bundle eval, before any
composition script runs) and prefer the stamp when hiding/restoring.
Also re-sync the grading canvas when the source's inline geometry mutates
(rAF-throttled style observer): a studio drag moves the source via its
transform, which fires no media event, so the visible canvas froze in
place until the next seek.
* fix(hooks): scope pre-commit build check to the actual target repo
The PreToolUse hook matched any Bash command containing "git commit" and
always ran this repo's bun build/lint/typecheck from its own process cwd,
even when the command targeted a different repo (e.g. a sibling worktree
reached via a leading `cd`). Resolve the real target directory from the
command text first, and skip silently for any repo that isn't this one.
* fix(lefthook): force-add already-tracked files under gitignored paths
The format hook's auto-restage (`git add {staged_files}`) exits non-zero
for any staged file that lives under a gitignored directory (e.g.
.claude/settings.json, tracked despite .claude/ being ignored for worktree
noise), silently aborting the whole commit even though the file was
already correctly staged.
* fix(producer): repoint stale puppeteer symlinks to the pinned 25.x install
packages/producer's tracked node_modules symlinks still pointed at
puppeteer@24.43.1, which no longer exists after a fresh install resolves
package.json's ^25.2.1 range to 25.3.0 — breaking the producer TypeScript
build with a missing puppeteer-core module error.
* fix(producer): stop tracking node_modules symlinks
packages/producer/node_modules was accidentally swept into a prior commit
despite the repo-wide node_modules/ gitignore rule, and its ~30 tracked
symlinks silently drift from whatever bun install actually resolves —
the exact cause of the stale puppeteer symlinks fixed earlier in this
branch. CI always runs bun install --frozen-lockfile before building, so
nothing depends on these being pre-committed.
* fix(telemetry): expose stalled render stages
* fix(telemetry): preserve capture data on terminal stage events
* fix(telemetry): fix calibration TDZ crash, tag encode/assemble, extend heartbeat cadence
capture_calibration referenced captureStageObservationData before its
declaration (later in the same scope), which would throw a ReferenceError
for any render hitting the calibration path. Hoist the closure and split
workerCount's declaration from its resolution so calibration can safely
read it as undefined before capture strategy resolves worker count.
Also address the two non-blocking review items: wire encode/assemble
stages through captureStageObservationData for consistent tagging, and
extend the heartbeat schedule to repeat every 120s after the initial
30/60/120s ramp instead of going dark on stalls beyond two minutes.
Two non-blocking review notes from Rames, both addressed:
1. Trial polarity inverted to OPT-IN: disableDeParallelRouterTrial →
enableDeParallelRouterTrial. renderLocal is exported, so any programmatic
consumer (future studio-server path, test harness, distributed runner)
previously inherited the trial and its process-wide env-var/module-latch
state without knowing to disable it — and concurrent invocation races
that state. Now only the CLI's own sequential call sites opt in (the
single top-level render, and batch at concurrency 1); everyone else gets
no trial by default. The doc comment names the sequential-invocation
assumption explicitly.
2. deSelfVerifyFallback semantic narrowing documented at both declarations
(RenderCaptureObservability + RenderPerfSummary.drawElement): since the
pinned-fallback retry was widened, the flag means verify-triggered
SPECIFICALLY — OOM/capture_error fallbacks report false with
deFallbackReason carrying the reason. Dashboards keyed on
de_self_verify_fallback=true as "any fallback fired" must migrate to
de_fallback_reason IS NOT NULL (also called out in the PR body for the
observability rebuild to pick up).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`hyperframes skills update` failed hard or looped forever once a skill was
retired/renamed upstream while still installed locally (hyperframes-media folded
into media-use; hyperframes-captions/compose/tts consolidated earlier). Two
paths dead-ended:
- Install: target selection could trust a stale local skills-manifest.json
(findRepoManifest) while `skills add` always installs from the canonical repo.
isCoreSkill matches the `hyperframes-` prefix, so a retired skill was forced
into the target set, `skills add` silently declined it (exit 0), and strict
verifyInstalled threw "Skill(s) still missing after install".
- Prune: upstream `skills remove` scans on-disk directories, so a lock entry
retired before it ever shipped a bundle has nothing to match — a silent
exit-0 no-op that never clears the lock, so detectRemoved re-flags it on
every run.
The stale-skills nudge compounded it: it fired even from `skills update` itself
(pointing users back at the failing command) and its count ignored the removed
bucket.
Resolve update targets against the canonical manifest (checkSkills({ canonical:
true })) so a retired skill is never targeted. Add pruneOrphanedLockEntries to
clear the orphaned lock entries the upstream remover can't (idempotent, so a
second run is a clean no-op). Exclude `skills` from the update-nudge gate and
thread the removed count through the nudge total.
The pipeline already carried FrameCheckOptions; only the flag was
boolean, which meant a pipeline caller tuning severity or seek points
would have them silently dropped — the two sides only agreed because
today's caller happens to match the defaults. Bare --frame-check keeps
the defaults; the value form mirrors --caption-zone's grammar, freezing
the contract before a release pins it.
Port validate's per-media-element clip audit into check's session
(clip_media_fit findings): an intrinsic duration meaningfully shorter
than the data-duration slot silently shortens the slot at render time,
and neither lint nor the runtime listeners can see it. A linter crash
now reports as check_lint_failure instead of masquerading as a runtime
failure. Finding-crop capture failures stay non-gating but emit a
stderr note and a telemetry error event so rollouts can measure the
second-session failure rate.
Five findings from a fifth (final scoped) max-effort review of the previous
commit, all local:
1. writeConfig now writes atomically (pid-suffixed temp file + renameSync —
rename within one directory is atomic on POSIX). This closes the real
hazard behind the review's torn-read finding: readConfig's corrupted-file
catch RESETS the config to defaults (telemetry re-enabled, anonymousId
rotated, trial fields wiped), so a concurrent reader catching a
non-atomic write mid-flight would silently destroy the user's config —
and the previous commit's per-render readConfigFresh() at the arm site
multiplied exposure to exactly that window. Verified against a real
filesystem, not just the mocked unit tests.
2. writeConfig now returns whether the write landed (errors still swallowed
— telemetry must never break the CLI). persistDeParallelRouterTrialFired
uses it to stop immediately on a genuine fs failure (retrying an
unwritable file is pointless) and reserve its retries for actual
concurrent clobbers, instead of 3 blind write attempts + 4 disk reads.
3. The persistence-failure console.warn is now !quiet-gated like every
other trial message — a quiet/batch-json render on an unwritable
~/.hyperframes no longer emits unexpected stderr that CI wrappers
asserting empty stderr would misread as a render failure. The in-process
latch already guarantees the safety behavior whether or not the warning
prints.
4. The arm site short-circuits on the in-process fired latch BEFORE the
fresh config read — post-fired batch rows no longer pay a per-row config
read + parse + shared-cache invalidation for an answer module state
already knows.
5. Replaced the new `as T` assertions in render.test.ts's config-state
factory with an explicitly typed vi.hoisted return (repo TypeScript
convention: no `as T`).
config.test.ts: node:fs mock gains renameSync (faithful to the new atomic
write); new test covers the success/failure return and asserts no temp file
survives a write.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Three root causes from a fourth max-effort review (15 raw findings deduped;
the synthesize step died on a session limit so they arrived unmerged):
1. The previous commit's telemetryEnabled fix was ineffective: the arm site
passed readConfig() — the process-lifetime cache — into
isDeParallelRouterTrialBlocked, making it exactly as stale as the
shouldTrack() memoization it claimed to bypass. A mid-batch
`hyperframes telemetry off` (or another process persisting fired=true)
was never observed. Now reads readConfigFresh() at the arm site; the
test mock previously hid this because readConfig/readConfigFresh were
behaviorally identical views over one shared object.
2. The verify-and-retry write loop double-counted a render whenever OUR
write landed but a concurrent writer advanced the file before our
verify read — the retry re-applied the increment on top (two renders
→ three counts), tripping the 25-render exposure cap early and
permanently killing the trial with less telemetry than the cap was
designed to allow. Reworked: the render COUNTER is written exactly
once, unverified (a lost increment under-counts by one — benign); only
the FIRED flag is verified and re-asserted, which is idempotent, so
retries can no longer corrupt anything
(persistDeParallelRouterTrialFired).
3. writeConfig swallows all fs errors, so on an unwritable ~/.hyperframes
a reverted outcome could never persist — the trial would re-arm and
re-fail on every subsequent render forever, silently. Added an
in-process fired latch (set at decision time, before persistence is
attempted) consulted by the blocked-check, plus a one-time console
warning when persistence exhausts its attempts. Later processes still
re-arm (disk is the only cross-process channel), but each process now
stops after at most one failure it couldn't record.
Test infrastructure fix enabling all of the above to be tested: the config
mock now models disk vs cache SEPARATELY (readConfig serves the cache,
readConfigFresh re-reads "disk", writeConfig updates both) with a
failWrites hook simulating the real writeConfig's silent error swallowing.
The old single-shared-object mock made cached-vs-fresh mis-routing and
retry iterations untestable by construction.
3 new regression tests: mid-batch opt-out observed through the cache;
fired flag re-asserted after a lost write WITHOUT re-counting the render;
unwritable-config latch blocking re-arm. 56 tests total across
render.test.ts + config.test.ts.
Not fixed (by design): the widened pinned-fallback retry paying a doubled
render on deterministic mid-stream failures (e.g. ENOSPC) — the accepted
tradeoff of the fallback design; cancellation and OOM are special-cased.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(core): escape NUL delimiters in HFMASK mask token and restore regex
Raw 0x00 bytes in the maskInertRegions token and restore regex made
timingCompiler.ts binary to git and shipped raw NULs into dist/cli.js.
Bun's transpiler (<= 1.3.11) corrupts raw NULs in regex literals into
literal backslash-uFFFD text, so restore never matched: every masked
<style>/<script> region was dropped, the player never initialized, and
bunx renders produced blank white frames showing HFMASK tokens.
Use \u0000 escapes instead, which survive any transpile layer, and add
a byte-level regression test (behavior is identical under Node, so only
a byte check catches this).
Fixes the first half of #2139.
* fix(cli): use NTFS junctions for studio project links on Windows
linkProjectIntoStudioData called symlinkSync(dir, path, "dir"), which
needs Developer Mode or elevation on Windows, so preview and dev in
local-studio mode died with EPERM for default-configured users.
Junctions need no privilege, work for directories, and keep the live
write-back the studio depends on (a copy fallback would decouple the
studio from the real project). Covers both preview and dev, which share
the helper.
Fixes the second half of #2139.
Running the CLI from source (tsx dev script, as CI's smoke job does)
transpiles with keepNames, which rewrites named inner functions in
serialized page closures into __name(...) calls — a helper that exists
in the Node bundle but not in the browser realm. The unified seek
closure was the first validate-path page function with named inner
functions, so 'hyperframes validate' threw '__name is not defined' in
CI while the dist build worked. Every session opener now installs a
no-op __name shim via evaluateOnNewDocument before any page script
runs, immunizing all serialized closures regardless of build mode.
Three filters keep the escalated contrast gate honest, each surfaced by
running check across the registry examples:
- data-layout-ignore (the layout audit's existing decorative opt-out)
now also excludes set-dressing text from the contrast audit — one
vocabulary for 'not copy a viewer must read'.
- Text that has (nearly) left the canvas is skipped: sampling a clamped
off-canvas box reads border pixels and produced the classic false
white-on-white (a cursor exiting the frame).
- Contrast failures follow the same persistence rule as layout findings:
observed at a single sample of a multi-sample sweep demotes to
warning; held failures gate.
Example fixes the sweep exposed: motion-blur's 21 deliberately-dim rail
labels are marked decorative (its vivid labels pass on their own);
nyt-graph's subtitle and source-note adopt the gate's suggested
compliant gray; decision-tree's declared duration drops 15s to 10s —
its content ends at ~9.5s and every render shipped a blank white tail.
Vance Ingalls