zenstory-ai/video-recap-skills / skills/video-recap
video-recap
Verified把一段或多集视频做成带画框字幕的中文解说成片或纯原声高光短片。用户提供 .mp4 / .mov / .mkv / .webm,要求做影视解说、短剧二创、加旁白配音、跨集混剪或导出剪映草稿时使用;串联完成看片识台词、定主线与留原声、按台词停顿切镜头、写词检查、TTS 配音与混音加字幕。触发词:视频解说、影视解说、视频旁白、生成解说、视频 recap、video recap、voiceover、narration、recap。
scripts/briefing/context.py
View on GitHub"""Load and format background research and moderation context for the brief."""
import re
def _scene_number(scene_id):
"""1-based scene number for every human-facing label in the brief (JSON keeps the
0-based `scene_id`); a scene split across kept clips ("3.1") reads "4 第2段"."""
base, _, part = str(scene_id).partition(".")
return f"{int(base) + 1} 第{int(part) + 1}段" if part else str(int(base) + 1)
def _clip_text(text, limit):
"""Collapse whitespace and clip. Research fields are optional, so None clips to ""."""
return re.sub(r"\s+", " ", str(text or "")).strip()[:limit]
def _as_list(value):
if isinstance(value, list):
return value
return [value] if value else []
def normalize_background_research(research):
"""Coerce agent-written background_research.json into the shapes the formatters read.
Agents write bare strings where the research guide documents objects (plot_arcs,
cultural_notes) or lists (aliases, relationships); non-object character entries drop.
"""
characters = research.get("characters")
details = research.get("character_details")
return {
"synopsis": research.get("synopsis"),
"worldbuilding": research.get("worldbuilding"),
"episode_context": research.get("episode_context"),
"characters": characters if isinstance(characters, dict) else {},
"character_details": {
name: {
"aliases": _as_list(info.get("aliases")),
"role": info.get("role"),
"relationships": _as_list(info.get("relationships")),
}
for name, info in (details.items() if isinstance(details, dict) else ())
if isinstance(info, dict)
},
"plot_arcs": [
arc if isinstance(arc, dict) else {"description": arc}
for arc in _as_list(research.get("plot_arcs"))
],
"cultural_notes": [
note if isinstance(note, dict) else {"item": note}
for note in _as_list(research.get("cultural_notes"))
],
}
def _format_background_research(research, limit=1800):
"""Render background_research.json into a bounded Chinese story-context brief section."""
if not research:
return []
research = normalize_background_research(research)
lines = [
"## 故事背景资料(来自 background_research.json)",
"",
"背景调研仅供对照(`context-only`):只用于核对人名、别名、阵营设定与前情背景;未经本片画面或 ASR 证实的关系与后续剧情,不要直接写成当前画面的既成事实,也不要提前剧透。",
"",
]
for key, label in (
("synopsis", "剧情梗概"),
("worldbuilding", "世界观设定"),
("episode_context", "本集前情"),
):
value = _clip_text(research[key], 500)
if value:
lines.append(f"- {label}: {value}")
characters = research["characters"]
if characters:
lines.append("- 主要角色:")
for name, desc in list(characters.items())[:12]:
lines.append(f" - {_clip_text(name, 60)}: {_clip_text(desc, 160)}")
details = research["character_details"]
if details:
lines.append("- 角色详情:")
for name, info in list(details.items())[:8]:
bits = []
aliases = [_clip_text(alias, 40) for alias in info["aliases"][:4]]
if aliases:
bits.append("别名 " + "/".join(aliases))
role = _clip_text(info["role"], 80)
if role:
bits.append(role)
rels = [_clip_text(rel, 80) for rel in info["relationships"][:4]]
if rels:
bits.append(";".join(rels))
if bits:
lines.append(f" - {_clip_text(name, 60)}: {'; '.join(bits)}")
arcs = research["plot_arcs"]
if arcs:
lines.append("- 故事线索:")
for arc in arcs[:8]:
name = _clip_text(arc.get("name"), 80)
desc = _clip_text(arc.get("description"), 180)
status = _clip_text(arc.get("status"), 40)
if name or desc:
text = f"{name}: {desc}" if name and desc else name or desc
tail = f" [{status}]" if status else ""
lines.append(f" - {text}{tail}")
notes = research["cultural_notes"]
if notes:
lines.append("- 专有设定:")
for note in notes[:6]:
item = _clip_text(note.get("item"), 80)
expl = _clip_text(note.get("explanation"), 160)
if item and expl:
lines.append(f" - {item}: {expl}")
elif item or expl:
lines.append(f" - {item or expl}")
lines.extend(
[
"",
"写解说时优先使用角色真名或具体身份关系,避免对主要人物使用「黑衣男子」「白发女子」等泛称。",
"",
]
)
text = "\n".join(lines)
if len(text) <= limit:
return lines
clipped = text[:limit].rsplit("\n", 1)[0].rstrip()
return [
*clipped.splitlines(),
"",
"[背景资料较长,已截断以保留画面与 ASR 证据空间]",
"",
]
def _format_moderation_refusals(scenes_analysis):
"""One brief line naming the scenes MiMo moderation refused, so blank descriptions read as
refusals (vlm.analyze_scenes sets analysis_status), not as empty footage."""
refused = [
scene for scene in scenes_analysis or []
if isinstance(scene, dict) and scene.get("analysis_status") == "moderation_refused"
]
if not refused:
return []
labels = ", ".join(_scene_number(scene["scene_id"]) for scene in refused[:12])
more = f" …(+{len(refused) - 12})" if len(refused) > 12 else ""
return [
f"- 画面审核拒答场景: {len(refused)}/{len(scenes_analysis)}(场景 {labels}{more})— "
"VLM 请求触发内容审核而未返回描述与帧动作,并非空镜头;请结合 ASR 对白、原片硬字幕或背景调研核实,cut 模式下如无把握可跳过。",
"",
]
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |