← Back to TeamOpenClaw.com Get the AI Team Guide
Agent profile

Script Writing Agent

The Script Writing Agent takes approved topics from the Content Strategy Agent and ships camera-ready scripts in both short (60–90 seconds) and long (6–8 minutes) formats. Hooks, talking points, B-roll prompts, CTA language—it’s all drafted in your voice so you can hit record without staring at a blank doc.

Primary outputShort + long video scripts (Markdown)
TriggerManual or auto when topic status = “ready”
Human inputVoice notes (optional), final approval

What You Get

Each script package includes an intro hook tied to current market data, three supporting beats (story, stat, social proof), transitions, CTA copy, and notes for B-roll/cutaways. Scripts are stored in /content/scripts/<topic-id> so your editor or ISA can grab them instantly.

Inputs

  • Topic JSON from Content Strategy Agent
  • Research Agent metrics (auto-linked)
  • Your voice samples or tone guidelines
  • Optional: audio note or Loom if you want extra nuance

Outputs

  • short.md (60–90 seconds)
  • long.md (6–8 minutes)
  • CTA variations (organic + lead magnet)
  • Shot list + B-roll suggestions

Hand-offs

  • Social Media Agent pulls the hooks/CTAs
  • Follow-Up Agent references CTAs for nurture
  • Mission Control logs the script link per topic

Implementation Checklist

  1. Attach tone: Drop a sample video transcript or text snippet into life/projects/solo-agent-mission-control/prompts/script-tone.md. The prompt in Chapter 7 references it automatically.
  2. Map directories: Store scripts under /mission-control/public/scripts/<topic-id>/. The agent creates the folders if needed.
  3. Trigger logic: Add a Mission Control automation: “When topic status = ready → run script agent.” You can also kick it off manually with npm run agent:script.
  4. Approval loop: After each script, the agent @mentions you in Mission Control. Leave edits inline or hit approve and it forwards the files to your editor/ISA.

Best practices

  • Keep hook length < 20 seconds—agent enforces it if you specify
  • Feed the agent your dominance/credibility points quarterly
  • Use the optional brandFacts array for testimonials or awards

Quality guardrails

  • Links to original stat + date for every number mentioned
  • Automatic “compliance check” tag if you mention rates
  • Fallback CTA inserted if pipeline offers are marked inactive

Get the Script Writing Agent Prompt + Workflow

Chapter 7 includes the full Claude/GPT prompt chain, directory structure, and Mission Control automation so this agent always mirrors your voice. Install it once and you’ll never type another “Hey y’all, it’s Monday…” cold open again.

Get the AI Team Guide