Skip to main content
skill CX Codex

extracting-structure — Claude Code skill: chapters, bug reports, hook

Pulls structure out of a video you already watched: timestamped chapters, a bug report with the exact frame, a scored hook analysis.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add oxbshw/watch-skill --skill extracting-structure -a claude-code
Open on GitHub

What extracting-structure does

The skill answers from the persistent index of a video that has already been watched, so extraction is deterministic and costs no extra model calls. Watching comes first — that part belongs to the watching-videos skill.

Three extractors. watch-skill extract chapters <video_id-or-source> builds titled chapters with start and end timestamps from scene changes plus topic shifts in the transcript. watch-skill extract bug-report <video_id-or-source> pins down the failure: timestamp, the exact frame, the on-screen error text as OCR read it, and the steps leading up to it. When there is no error signal it returns found: false honestly, and that should be reported as no error found in the recording rather than as a failure. watch-skill extract hook <video_id-or-source> [--seconds 15] scores the opening seconds on four measured axes: attention trigger in the opening line, speech pacing, visual change rate, on-screen text. Each axis comes with a critique, plus a 0–100 score and a verdict.

The skill is limited to Bash and Read. For hook analysis, hand the user the per-axis critiques — the number on its own explains nothing.

What it helps with

  • build a chapter list with timestamps for a long tutorial
  • locate the exact frame where an error shows up in a screen recording
  • turn a screen recording into a fileable bug report
  • score an intro on pacing and visual change rate
  • split a lecture or meeting into topic segments

Example prompts for the agent

  • make chapters for this video
  • where does the bug appear in this recording
  • turn this screen recording into a bug report

How to install extracting-structure

1Skills CLI (Vercel)

npx skills add oxbshw/watch-skill --skill extracting-structure -a claude-code

2Manually: put the files in place

~/.claude/skills/extracting-structure/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/extracting-structure/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about extracting-structure

How do I install extracting-structure in Claude?

Run npx skills add oxbshw/watch-skill --skill extracting-structure -a claude-code. The skill lands at ~/.claude/skills/extracting-structure/SKILL.md.

What does extracting-structure do?

It extracts structure from an already-watched video: timestamped chapters, a bug report with the exact frame and error text, and a hook analysis across four axes with a 0–100 score.

Why did the command return found: false?

It means the recording holds no error signal. That is an honest result, not a failure — report it as no error found in the recording.

Source text

SKILL.md · MIT
---
name: extracting-structure
version: "1.4.3"
description: The user wants structure pulled out of a watched video — "make chapters for this video", "where does the bug appear in this recording", "turn this screen recording into a bug report", "how strong is my intro/hook". Use this for deterministic extraction from the index — chapters with timestamps, a fileable bug report with the exact frame, or a scored hook analysis.
license: MIT
user-invocable: true
allowed-tools: Bash, Read
---

# Extracting structure

Three extractors turn an already-watched video into something structured.
All of them answer straight from the persistent index — deterministic, no
extra model calls. The video must be watched first (see `watching-videos`).

## Chapters

```bash
watch-skill extract chapters <video_id-or-source>
```

Titled chapters with start/end timestamps, built from scene changes plus
transcript topic shifts. Good for long tutorials, lectures, meetings —
give the user the chapter list with timestamps, not a prose summary.

## Bug report from a screen recording

```bash
watch-skill extract bug-report <video_id-or-source>
```

Pinpoints where an error appears: timestamp, the exact frame, the
on-screen error text as OCR read it, and the steps that led up to it
(from the transcript/actions preceding the failure). Returns
`found: false` honestly when no error signal exists — report that as
"no error found in the recording", not as an error.

Paste the output into the user's issue tracker format when they ask for
a fileable report; the frame path is the attachment.

## Hook analysis (creators)

```bash
watch-skill extract hook <video_id-or-source> [--seconds 15]
```

Scores the opening seconds on four measured axes — attention trigger in
the opening line, speech pacing, visual change rate, on-screen text —
each with an actionable critique, plus a 0–100 score and verdict. Give
the user the per-axis critiques; the number alone helps nobody.

© oxbshw · MIT · oxbshw/watch-skill

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set11

watch watch examines linked or local videos, extracts frames and speech, then answers questions from an index. Use it to review a recording or check your own interface. CX Codex +1 583installs watching-videos Processes local video files and links through scene-aware frame extraction, OCR, and speech transcription. It lets the agent answer questions using verified timestamps. CX Codex +1 525installs video-memory A persistent index of every video the agent has analyzed on this machine: frames, OCR, transcripts and past answers. Useful when you need to recall what was in a recording. CX Codex +1 510installs asking-with-evidence Answers questions about previously indexed videos using persistent transcripts, OCR, and timestamps. It prevents re-watching and reports clear evidence or honest refusals. CX Codex +1 506installs the-loop Records how your UI, animation or game actually behaves, judges the recording against plain-language criteria, and iterates until it passes. CX Codex +1 501installs recovering-from-errors This skill reads a Watch Skill error code, runs watch-skill doctor, applies the stated fix and retries the failed command once. CX Codex +1 498installs learning-from-mistakes Turns your corrections of video answers into stored lessons, so later answers on similar questions get better. Also shows what has been saved. CX Codex +1 496installs configuring-vision Configures vision providers for Watch Skill: connects OpenAI, Anthropic, Gemini, OpenRouter, or local Ollama without locking to a single vendor. CX Codex +1 496installs

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor