技能 2026-06-23 12:25 UTC
speakerv0.8.0
版本【v0.8.0】由提交人 AI272 提交,主要更新类型为:功能特性新增与扩展。提交说明:Release v0.8: pacing model, compact pipeline, concise output。为项目引入了新的能力与模块支持。
Commit提交信息原文
Release v0.8: pacing model, compact pipeline, concise output - Add pause-aware pacing model and per-slide word budget (EN ~110 wpm, ZH ~165 chars/min); fixes 15-min scripts overrunning to ~30 min. - read_slides.py: add --mode compact for smaller intermediate JSON. - visual_inventory.py: add --ocr-scope image-regions; byte-safe OCR decoding and symlink path resolution. - vision_review.py: add --format compact (hoist shared prompt/schema). - write_display_docx.py: timing table with words/budget/pauses + TOTAL. - SKILL.md: glossary on/off toggle; final delivery is summary + paths. - Add LOGIC.md design doc; bump README/README.zh to v0.8 with notes. - Rebuild package as speaker-v8.skill (replaces speaker-v7.skill).