Agent-Skills.md

Agent Skills: Generate Office Hours Write-up

Generate a structured Q&A write-up from a weekly Python + AI Office Hours recording. Use when asked to create an office-hours write-up, extract questions and answers, or process a recording for a specific date.

UncategorizedID: pamelafox/office-hours-writeups/generate-writeup
181

Install this agent skill to your local

pnpm dlx add-skill https://github.com/pamelafox/office-hours-writeups/tree/HEAD/.agents/skills/generate-writeup

Skill Files

Browse the full folder contents for generate-writeup.

Download Skill

Loading file tree…

.agents/skills/generate-writeup/SKILL.md

Skill Metadata

Name
generate-writeup
Description
"Generate a structured Q&A write-up from a weekly Python + AI Office Hours recording. Use when asked to create an office-hours write-up, extract questions and answers, or process a recording for a specific date."

Generate Office Hours Write-up

Create an easy-to-read, searchable Q&A write-up for a weekly Python + AI Office Hours recording.

Inputs

  • The office-hours date, normally supplied as YYYY-MM-DD
  • The YouTube URL, either supplied by the user or available in that date's existing files

If the date is not explicit, infer it only when the target date is unambiguous from the current context or existing office-hours directory. Otherwise, ask for it.

Procedure

  1. Use the youtube-transcript skill to fetch the recording transcript with timestamps. Save the unmodified result as office-hours/YYYY_MM_DD/transcript.md.
  2. Use the youtube-live-chat skill to fetch the live chat. Save it as office-hours/YYYY_MM_DD/live_chat.md. If YouTube has no chat replay, check raw.md for pasted Discord chat and use that as supplemental context.
  3. Read the transcript and chat, identify the main questions and answers, distinguish follow-up questions from new topics, and identify substantive discussions, demos, and announcements that are not questions.
  4. Write office-hours/YYYY_MM_DD/questions_answers.md using the format below.
  5. Check that every heading has the correct timestamp link and that factual claims and shared links are grounded in the transcript or chat.

Output Rules

  • Begin with a descriptive title such as # July 21, 2026 Office Hours Q&A.
  • Use ## headings for main questions and ### headings for follow-up questions.
  • When a substantive segment is not a question, use a descriptive ## Discussion: ..., ## Demo: ..., or ## Announcement: ... heading in the appropriate place in the session chronology. Subtopics within a discussion or demo may use descriptive ### headings.
  • Put a linked video timestamp immediately below every ## or ### heading.
  • Summarize answers clearly rather than reproducing the transcript verbatim.
  • Add inline links to natural words or phrases in the answer text.
  • Add a Links shared: list only for links that do not have an obvious inline anchor.
  • Do not include upcoming events as a write-up section.
  • Preserve meaningful caveats, uncertainty, and distinctions from the spoken answer.

Format

# Month D, YYYY Office Hours Q&A

## Question 1?

📹 [12:34](https://youtube.com/watch?v=VIDEO_ID&t=754)

Answer to question 1 with a natural [inline link](https://example.com).

Links shared:

* [Link without a natural inline anchor](https://example.com)

### Follow-up question 1a?

📹 [15:20](https://youtube.com/watch?v=VIDEO_ID&t=920)

Answer to the follow-up question.

## Question 2?

📹 [18:45](https://youtube.com/watch?v=VIDEO_ID&t=1125)

Answer to question 2.

## Demo: Using a new agent tool

📹 [24:10](https://youtube.com/watch?v=VIDEO_ID&t=1450)

Summary of what the demo showed.

## Discussion: Agent security practices

📹 [31:05](https://youtube.com/watch?v=VIDEO_ID&t=1865)

Summary of the discussion, including uncertainty and differing viewpoints.

## Announcement: New Python release

📹 [45:30](https://youtube.com/watch?v=VIDEO_ID&t=2730)

Summary of the announcement.