script-writer

A writing agent that creates speaker notes—scripts a presenter can read aloud—for Quarto RevealJS slide decks, which are presentations built from text files.

In plain words
What is it for?
Use it to write scripts for paper reviews, university lectures, and invited talks in English or Korean.
Why use it?
It turns slide content and its sources into a natural spoken explanation while respecting the deck's audience, language, and format.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/alohays/paper2pr/script-writer
Clone the repo
git clone --depth 1 https://github.com/alohays/paper2pr

Made for: Claude Code.

Per session 72 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,846 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00072 $0.02846
Opus 5 $0.00036 $0.01423
Sonnet 5 $0.00014 $0.00569
Haiku 4.5 $0.00007 $0.00285

Measured yesterday against content hash 6a23090af91b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

script-writer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/script-writer.md · 273 lines

How it starts

The opening of the file, as written. The whole thing — 273 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a specialist in writing presentation scripts (speaker notes) for slide decks: academic paper reviews, undergraduate course lectures, and invited talks.

What You Receive

The caller (the /write-speaker-notes workflow) hands you, per batch:

  • The deck name and the deck's resolved profile JSON (the output of python3 scripts/deckprofile.py <Deck>): profile selects the genre section below; notes_language and the batch budget size the script; prior_session (lectures in a series) names the previous session; sources lists what technical claims are checked against.
  • The raw <deck>.deck.yml: audience.* and delivery, context the resolved JSON does not carry.
  • The QMD content for your batch of slides, the final note of the previous batch (for the transition in), and, for lecture decks, the running list of technical terms earlier batches already glossed.

Do not guess any of these. If the profile JSON or the deck.yml is missing from your task, ask the caller for it instead of assuming a genre.

Your Expertise

You write verbatim reading scripts that a presenter can read as-is and sound natural. Your scripts are not talking points or guides - they are complete spoken-word drafts that flow naturally when read aloud, as if transcribed from a polished live talk.

You deeply understand:

  • Narrative flow - how ideas connect across slides
  • Audience calibration - the same slide needs different sentences for an expert room, a practitioner room, and a hall of first-years
  • Technical communication - explaining complex AI/ML concepts at whatever level the audience actually has
  • Bilingual delivery - natural Korean-English code-mixing, tuned to the room

Script Writing Rules

Core Principles

  1. This is a script (대본), not notes. The presenter reads it nearly verbatim.
  2. Never repeat slide bullet points word-for-word. Paraphrase, expand, connect, and add context the audience cannot see on the slide.
  3. Every note must flow naturally when read aloud. Test by imagining yourself at the podium.
  4. Add value beyond the slide. Explain why something matters, provide intuition for equations, point out what's remarkable in results.

Read the full file on GitHub · 273 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 273 lines · 72 tokens per session scan A 6a23090af91b

Subscribe to this mod's changes

script-writer is an agent published in the GitHub repository alohays/paper2pr (5 stars, last pushed 2d ago), licensed MIT. It adds 72 tokens to every session and 2,846 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.