Skip to content

brief(phase4) + scaffold: PR-5 phase 0 follow-up - #25

Merged
sidorovanthon merged 1 commit into
mainfrom
pr5/phase0-followup
May 1, 2026
Merged

sidorovanthon merged 1 commit into
mainfrom
pr5/phase0-followup

Conversation

@sidorovanthon

Copy link
Copy Markdown
Owner

Summary

  • Brief: Sub-composition split block flipped from "strong recommendation" to "BLOCKED pending upstream fix" — canon-blessed pattern produces black renders in HF 0.4.41 per Phase 0 findings (PR docs(spec): phase 0 investigation findings #24).
  • Q3 scaffold workaround: verified empty by grep across scripts/ and .claude/commands/ — no scaffold patch needed (YAGNI).
  • Filed 2 upstream HF issues with bare repros.

Upstream issues filed

Canon respect

Brief change does not edit canon. The block now explicitly cites the broken canon-blessed pattern + the upstream issue tracking the fix. Once upstream merges + we bump HF version, this block reverts to the imperative "MUST split for ≥3 beats" form (tracked in memory).

Test plan

  • Brief reads coherently with PR-2 preamble + PR-1 diagnostics block.
  • Both upstream issue URLs resolve.
  • Step 9 memory entries cite the upstream issue URLs (not just our PR).

🤖 Generated with Claude Code

Phase 0 bare-repro (PR #24) confirmed three upstream HF 0.4.41
bugs. PR-5 lands the orchestrator-side reactions:

1. Brief: sub-composition split block changed from "strong
   recommendation / SHOULD" to "BLOCKED pending upstream fix".
   The canon-blessed <template> + data-composition-src pattern
   produces black renders in HF 0.4.41; do not split beats into
   sub-comps until upstream loader unwraps <template>.content.
   Tracks upstream issue
   heygen-com/hyperframes#589.

2. Q3 scaffold workaround: verified via grep — scripts/ and
   .claude/commands/ do not emit the literal "repeat:-1" in
   comments, so no scaffold patch needed (YAGNI).

Filed two upstream issues:
- heygen-com/hyperframes#589 — sub-comp
  loader / compositions CLI vs <template>
- heygen-com/hyperframes#590 — lint
  regex false-positive on JS comments

Step 9 (held memory writes) follows.

Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>
@sidorovanthon
sidorovanthon merged commit d94eaa1 into main May 1, 2026
@sidorovanthon
sidorovanthon deleted the pr5/phase0-followup branch May 1, 2026 10:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant