Agent skill

watch-summary

Generate a video-ready narration script summarizing the latest chat session

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/watch-summary

SKILL.md

/watch-summary - Session Summary for Video Narration

Generate a concise, spoken-word summary of the latest chat session — formatted as a prompt for AI video/audio generation. Watch and hear your session instead of reading it.

Usage

/watch-summary              # Summarize this session
/watch-summary <agent>      # Summarize from a specific agent's perspective

Instructions

Step 1: Identify What Happened

Review the current conversation and extract:

  • Changes made — files created, edited, deleted
  • Decisions — what was chosen and why
  • Ideas discussed — brainstormed but not yet implemented
  • Problems solved — bugs fixed, blockers cleared
  • New tools/integrations — anything added to the system
  • Status updates — what moved forward, what's next

Step 2: Write the Video Script

File location: Place the summary file inside the project folder that was the focus of the session:

  • Dashboard/Work_Space/[ProjectFolder]/Watch_Summary_[YYYY-MM-DD].md
  • Example: If the session was about AutoMates-Brand → Dashboard/Work_Space/AutoMates-Brand/Watch_Summary_2026-03-11.md
  • Example: If the session was about Sunny → Dashboard/Work_Space/Sunny/Watch_Summary_2026-03-11.md
  • If the session spanned multiple projects, pick the primary one.
  • If no specific project folder applies (e.g. system-wide work), use Dashboard/Work_Space/Watch_Summary_[YYYY-MM-DD].md as a standalone file.

Do NOT create a Watch_Summaries/ folder. Folders in Work_Space are for projects only.

Use this format:

markdown
# Session Summary — [YYYY-MM-DD]

**Agent:** [who was active]
**Duration:** [approximate, based on conversation length]
**Generated:** [timestamp]

---

## Video Narration Script

[Write a natural, conversational narration script — as if briefing the user over coffee. Use short sentences. No jargon dumps. Speak in second person ("you", "your team"). Keep it under 2 minutes when read aloud (~300 words max).]

Structure the narration:
1. **Opening** (1-2 sentences) — What this session was about
2. **Key changes** (3-5 bullets narrated naturally) — What actually changed
3. **Decisions made** (if any) — What was decided and why
4. **What's next** (1-2 sentences) — Where things stand now

---

## Quick Reference

### Changes
| What | Where |
|------|-------|
| [change description] | [file/location] |

### Decisions
- [decision]: [reasoning]

### Ideas (Not Yet Implemented)
- [idea]

### Next Steps
- [next action]

---

*Generated by /watch-summary — paste the Video Narration Script into your AI video tool.*

Step 3: Writing Rules

  • Conversational tone — this will be spoken aloud, not read on screen
  • Second person — "You integrated Playwright CLI" not "Playwright CLI was integrated"
  • No filler — every sentence should carry information
  • ~300 words max for the narration script — aim for a 1-2 minute video
  • Be specific — "updated 3 Knowledge files" not "made some updates"
  • Skip internal mechanics — the user doesn't need to hear "I read Checkpoint.md then appended to Lessons.md"
  • Lead with impact — what matters most comes first

Step 4: Report

=== WATCH SUMMARY ===
Session: [date]
Agent:   [who was active]
Script:  Dashboard/Work_Space/[ProjectFolder]/Watch_Summary_[YYYY-MM-DD].md
Words:   [~word count of narration]
Est:     [~minutes when read aloud]
=====================

Print the narration script directly in the chat as well, so the user can copy-paste immediately.


Paste the narration script into any AI video/audio tool (NotebookLM, HeyGen, ElevenLabs, etc.) to generate your session recap.

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results