cut-mistakes

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Cut Mistakes & Repeats (Pipeline Agent 2)

剪切错误与重复内容（Pipeline Agent 2）

Second step of the automated edit, run after cut-silences. It removes the things a human editor cuts on a second pass: stutters ("the the", "I- I-"), immediately repeated words, false starts (an abandoned phrase that restarts), and retakes (a line re-recorded — keep the clean take, drop the botched one).

Deciding what counts as a mistake is judgment, not a formula — emphatic repetition ("never, never") and rhetorical doubling ("the first piece is, is this…") look identical to a stutter mechanically. So this agent is review-gated by default: it surfaces candidates with full context; the agent (and the user) decide; only approved cuts are rendered.

自动化剪辑的第二步，在 cut-silences 之后运行。它会移除人工剪辑师在第二轮剪辑中通常会删除的内容：结巴（“the the”“I- I-”）、紧邻的重复词、说错后重说（中途放弃某个短语并重新开始），以及重录内容（同一句台词被重新录制——保留干净版本，删掉录坏的版本）。

判断什么算作错误，需要的是 判断而非套用公式——强调性重复（“never, never”）和修辞性复沓（“the first piece is, is this…”）在机械规则看来与结巴完全一样。因此，此 Agent 默认采用 review-gated：它会呈现候选片段及其完整上下文；由 Agent（和用户）决定；仅渲染获批的剪辑。

When to use

何时使用

"cut the mistakes", "remove stutters / repeats", "clean up the flubs", "keep the best take", "cut false starts"
As the second stage of the master edit workflow, on cut-silences' output.

“剪掉错误”“移除结巴/重复”“清理口误”“保留最佳一条”“剪掉说错后重说的内容”
作为主剪辑工作流的第二阶段，处理 cut-silences 的输出。

The three-part flow

三段式流程

1. Find candidates (mechanical)

1. 查找候选片段（机械规则）

bash

node .agents/skills/cut-mistakes/scripts/find-cut-candidates.mjs \
  <silence-transcript.json> --out-dir video-projects/<slug>/assets

Input is normally Agent 1's

<stem>.silence-transcript.json

, so cuts land on the already-silenced timeline. Writes:

```
<stem>.cut-candidates.json
```
— structured candidates (type, confidence, proposed
```
cut
```
range,
```
removes
```
/
```
keeps
```
text, context, recommendation)
```
<stem>.cut-candidates.md
```
— readable proposal

Candidate types:

stutter

(immediate word repeat),

retake

(duplicate/near-duplicate segment via Jaccard similarity),

false_start

(short abandoned phrase that restarts).

bash

node .agents/skills/cut-mistakes/scripts/find-cut-candidates.mjs \
  <silence-transcript.json> --out-dir video-projects/<slug>/assets

输入通常是 Agent 1 生成的

<stem>.silence-transcript.json

，这样剪辑点就会落在已完成静音处理的时间线上。生成：

```
<stem>.cut-candidates.json
```
——结构化候选片段（类型、置信度、建议的
```
cut
```
范围、要
```
removes
```
/
```
keeps
```
的文本、上下文、建议）
```
<stem>.cut-candidates.md
```
——便于阅读的剪辑提案

候选类型：

stutter

（紧邻的词重复）、

retake

（通过 Jaccard similarity 识别的重复/近似重复片段）、

false_start

（中途放弃并重新开始的简短片段）。

2. Review (the gate — REQUIRED)

2. 审核（必需的门控步骤）

Read each candidate in context and decide keep-or-cut. The mechanical detector cannot tell intentional emphasis/rhetoric from a real flub, so do not trust

recommend

blindly — read

removes

context

. Common false positives: emphatic repetition ("it's never, never hands-off"), copula-then-question ("the piece is, is this…"), listing.

Present the candidates to the user with a recommendation per item; collect approvals. Write the approved cuts to a file:

json

{ "cuts": [ { "start": 100.93, "end": 101.06, "reason": "I- false start" }, ... ] }

(

start

end

in seconds on the input transcript/video timeline. You can widen a candidate's range — e.g. for a retake, cut from the botched take's start to the clean restart.)

结合上下文逐个阅读候选片段，并决定保留还是剪掉。机械检测器无法区分有意强调/修辞与真正的口误，因此不要盲目信任

recommend

——应阅读

removes

context

。常见误报包括：强调性重复（“it's never, never hands-off”）、系动词后接疑问句（“the piece is, is this…”）、列举表达。

向用户呈现候选片段，并为每一项提供建议；收集批准结果。将获批的剪辑写入文件：

json

{ "cuts": [ { "start": 100.93, "end": 101.06, "reason": "I- false start" }, ... ] }

（

start

end

以秒为单位，基于输入 transcript/video 的时间线。可以扩大候选片段的范围——例如，对于重录内容，可以从录坏版本的起点剪切到干净重录的起点。）

3. Apply approved cuts

3. 应用获批的剪辑

bash

node .agents/skills/cut-mistakes/scripts/apply-cuts.mjs \
  <silence-transcript.json> --cuts approved-cuts.json \
  --video video-projects/<slug>/assets/edited-silenced.mp4 \
  --output video-projects/<slug>/assets/edited-clean.mp4 --apply

Writes the EDL, a re-timed

<stem>.mistakes-transcript.json

(feeds the motion-graphics agent +

validate-beat-sync.mjs

), a decisions log, and — with

--apply

— the cut video (ffmpeg trim+concat, A/V in sync). Drop

--apply

for a dry run.

bash

node .agents/skills/cut-mistakes/scripts/apply-cuts.mjs \
  <silence-transcript.json> --cuts approved-cuts.json \
  --video video-projects/<slug>/assets/edited-silenced.mp4 \
  --output video-projects/<slug>/assets/edited-clean.mp4 --apply

生成 EDL、重新计时的

<stem>.mistakes-transcript.json

（供 motion-graphics Agent 和

validate-beat-sync.mjs

使用）、决策日志；使用

--apply

时，还会生成剪辑后的视频（通过 ffmpeg 执行 trim+concat，音视频保持同步）。移除

--apply

即可进行试运行（dry run）。

Reviewing the result

检查结果

Use the shared review tool to eyeball the cuts and spot-check boundaries:

bash

node scripts/build-edl-review.mjs <stem>.mistakes-edl.json \
  --original video-projects/<slug>/assets/edited-silenced.mp4 \
  --edited   video-projects/<slug>/assets/edited-clean.mp4 \
  --output   video-projects/<slug>/assets/mistakes-review.html
npx serve . -p 8080 -n

使用共享审核工具目视检查剪辑效果并抽查边界：

bash

node scripts/build-edl-review.mjs <stem>.mistakes-edl.json \
  --original video-projects/<slug>/assets/edited-silenced.mp4 \
  --edited   video-projects/<slug>/assets/edited-clean.mp4 \
  --output   video-projects/<slug>/assets/mistakes-review.html
npx serve . -p 8080 -n

Notes

注意事项

A very clean delivery may yield few or zero real cuts — that's a valid outcome; don't cut natural speech to hit a quota.
Cuts between two spoken words are hard joins. They're usually clean for stutters/false starts; for tighter audio a 20-30ms fade can be added later.
Hand the
```
mistakes-transcript.json
```
+ the clean video to Agent 3 (motion graphics / tiered cards).

如果表达非常干净，可能只会得到很少甚至零个实际剪辑点——这是有效结果；不要为了达到数量指标而剪掉自然表达。
在词语之间直接剪切会形成较生硬的衔接。对于结巴或说错后重说，通常不会有问题；若要让音频更紧凑，之后可以添加 20-30ms 的淡化（fade）。
将
```
mistakes-transcript.json
```
和剪辑后的视频交给 Agent 3（motion graphics / tiered cards）。