Kimi Agent Swarm — SKILL.md

---
name: "Kimi Agent Swarm"
description: "Use Kimi's Agent Swarm to run multiple AI tasks in parallel — scale research, content generation, and analysis across many agents simultaneously. For browser automation of kimi.com/agent-swarm."
version: "1.0.0"
author: "skynet"
category: "dev"
agents: ["claude-code"]
tags: ["kimi", "agent-swarm", "parallel", "browser-automation", "scaling"]
tools_required: ["chrome-mcp"]
---

# Kimi Agent Swarm

# Kimi Agent Swarm

Browser automation skill for kimi.com/agent-swarm — run multiple AI tasks in parallel.

## Navigation

Navigate to `kimi.com/agent-swarm` to access the Agent Swarm interface.

## Core Interface Elements

### Swarm Textbox
- The main textbox accepts task descriptions
- Describe a batch of work and Kimi distributes it across multiple agents
- Tasks run in parallel for maximum throughput

## How Agent Swarm Works

1. Describe your batch task in the textbox
2. Kimi breaks it down and distributes to multiple agents
3. Agents work in parallel on their assigned portions
4. Results are aggregated and returned

## Best Use Cases
- **Bulk research** — research many topics simultaneously
- **Parallel content generation** — create multiple pieces of content at once
- **Data analysis at scale** — analyze datasets across multiple agents
- **Mass summarization** — summarize many documents in parallel

## Factory Integration

Kimi Agent Swarm is particularly useful for the Factory system:
- A single Factory step could use Agent Swarm to parallelize work
- Great for steps that need to process many items independently
- Combine with other Factory steps for powerful pipelines

## Kimi Documentation
- Check Kimi docs for swarm API details and limits
- Agent Swarm may have rate limits on concurrent agents

## No Project System Warning

Kimi Agent Swarm has NO project system or persistent memory. Agents must include full context in every prompt:
- Who the user is and their account details
- What they are working on and why
- The complete goal and any relevant history
- All necessary URLs, credentials references, and prior results

Never assume the agent remembers anything from previous interactions.

## Running on Mac Minis

These browser automation tasks run on Mac Minis via mac-control on port 8200:
- **Vault**: 192.168.86.27:8200
- **Bots**: 192.168.86.50:8200
- **Jarvis**: 192.168.86.51:8200

Use the mac-control API to send browser automation commands. Example:
```
POST http://192.168.86.27:8200/execute
```