agentleFS
GoogleKnown publishergoogle-gemini/gemini-managed-agents-templates/ai-radio/skills/script-writing/SKILL.md

script-writing

Skill99 starsChanged 6 months agoApache-2.0View on GitHub

Generate the AI Talk Radio show script from research using the Interactions API.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Must include audio tags in square brackets (e.g., [sighs], [excitedly]).

    line 80

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Script Writing

Generate a naturalistic, multi-character radio show script directly via the LLM. The script features host Paul taking calls from people around the world, with advanced audio tags.

Embedded Script

python3 skills/script-writing/scripts/generate_script.py --workspace ./workspace --style debate

Arguments

ArgumentDefaultDescription
--workspaceworkspaceRoot workspace directory
--styledefaultShow format (see styles below)
--context""Additional tone/style notes inferred from user prompts (e.g., "emphasize the technical details", "make it sound like a late night show"). Applies to ALL styles. Keep brief. Do NOT use to specify stories/topics.

Styles

StyleCallersFormatUse when user says...
debate2 per topic, opposing viewsStructured back-and-forth, Paul moderates"opposing views", "debate", "both sides"
roundtable3-4, different anglesCollaborative discussion, callers build on each other"roundtable", "panel", "discussion"
interview1-2 with direct experienceQ&A format, Paul asks probing questions"interview", "deep dive", "expert"
explainer2-3, each explains an aspectTeach the audience, Paul asks clarifying questions"explain", "break it down", "what is"
defaultSame as debateDefault when no style specified(anything else)

Examples

# HN discussion with opposing views, applying context
python3 skills/script-writing/scripts/generate_script.py \
  --workspace ./workspace --style debate --context "make it sound like a late night show"

# Panel discussion about a GitHub repo with context notes
python3 skills/script-writing/scripts/generate_script.py \
  --workspace ./workspace --style roundtable --context "emphasize the performance improvements"

# Interview with an expert about a paper
python3 skills/script-writing/scripts/generate_script.py \
  --workspace ./workspace --style interview

# Explain a new technology
python3 skills/script-writing/scripts/generate_script.py \
  --workspace ./workspace --style explainer

What it does

  1. Reads all research from {workspace}/data/research/*.md.
  2. Builds a system prompt with the base format rules + style-specific caller/structure instructions.
  3. Calls Gemini via the Interactions API (client.interactions.create()).
  4. Saves the script to {workspace}/data/script.md.

Dependencies

  • google-genai (>= 2.0.0)

Output

  • Script File: {workspace}/data/script.md
  • Format: Plain text, each line starting with Speaker: dialogue

Cast

SpeakerRolePersonality
PaulHostProfessional but relatable British moderator
CallersCall-insAmateur, rough, natural — from various cities around the world

Script Format Rules

  • Every line: SpeakerName: [audio tag] Dialogue text
  • Must include audio tags in square brackets (e.g., [sighs], [excitedly]).
  • Callers should sound amateurish (include "uh", "like", natural pauses).
  • Short punchy sentences — spoken word, not prose.
  • ~450-500 words total (~3 minutes when spoken).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.