What is an Agent Pack?
An Agent Pack is a curated collection of actions that work together to solve a complete, end-to-end use case — things like researching a company, preparing for a meeting, or drafting a multi-touch outreach sequence. Each pack exposes two kinds of endpoints in the API Reference:- Orchestrated actions (e.g.
run_full_*,run_optimized_pipeline) that execute the whole workflow in one call. - Granular step actions you can mix, match, and customize in the Agent Builder.
Available packs
Company Research v2
Deep, cached company intelligence with per-section async research, tracking, and Q&A.
Competitive Brief
Generate a full competitive intelligence brief with one call.
Meeting Prep
Multi-phase meeting preparation with contact research and personalized sections.
Meeting Follow-Up
Analyze meeting transcripts and auto-generate coaching, follow-ups, and CRM notes.
Prospect Research
Triangulate prospect intelligence from LinkedIn, web, and company data.
Prospect Finder
Natural-language prospect discovery with ICP scoring.
Lead Qualifier
ICP-grounded lead scoring with Pursue / Nurture / Pass recommendations.
Outreach Drafter
Personalized email + LinkedIn message variants with personalization audit.
Sales Outreach Team
Outreach Strategist, multi-touch Sequence, and Cold Call Prep in one pack.
Content Marketing Team
AEO/SEO Research, Content Planner, Creator, and Audit.
Social Content Team
Plan, create, and analyze LinkedIn social content end-to-end.
YouTube Creator Team
Idea research, scripts, packaging, and cross-platform repurposing.
Industry Research
8-section market landscape brief with personalized implications.
Trends Research
6-section industry trends digest with email delivery.
HeyGen Avatar
Generate talking-avatar videos from a blog URL, topic, or script.
HeyGen Photo Avatar
Create custom photo avatars from uploaded or AI-generated images.
Instant Headshot
Studio-quality AI headshots from reference photos.
File Converter
Convert, merge, split, OCR, watermark, and encrypt files with CloudConvert.
OpenClaw
Talk to your personal OpenClaw instance over the OpenAI-compatible API.
When to use orchestrated vs granular actions
| Choose the… | When you want to… |
|---|---|
Orchestrated pipeline (run_full_*, run_optimized_pipeline) | Ship fast with default behavior, keep your agent small, or call the pack from another agent. |
| Granular steps | Customize intermediate outputs, branch on results, inject your own prompts/LLM, or cache expensive stages. |
| Render HTML action | Send the output as an email or display it in a human-readable report. |

