Agent Team Content

A multi-agent team for transforming conversations into multiple content formats. This team generates blog articles, social media posts, and presentations from a single input.

Agents

Agent	Description
blog	Creates engaging blog articles for Medium, Substack, and personal blogs
devto	Creates technical articles for the dev.to developer community
linkedin	Creates professional LinkedIn posts that drive engagement
twitter	Creates viral Twitter/X threads
marp	Creates Marp Markdown presentations with speaker notes
revealjs	Creates Reveal.js presentations with horizontal and vertical navigation

Workflow

The team uses a parallel workflow - all agents run independently on the same input conversation.

Project Structure

agent-team-content/
├── specs/
│   ├── agents/           # Agent definitions (*.md with YAML frontmatter)
│   │   ├── blog.md
│   │   ├── devto.md
│   │   ├── linkedin.md
│   │   ├── twitter.md
│   │   ├── marp.md
│   │   └── revealjs.md
│   ├── teams/            # Team workflow definitions
│   │   └── content-team.json
│   └── deployments/      # Deployment configurations
│       └── local.json
├── cmd/
│   └── content/          # Runtime CLI for generating content
├── .claude/
│   └── agents/           # Generated Claude Code agents
└── plugins/
    └── kiro/
        └── agents/       # Generated Kiro CLI agents

Generating Agent Files

Use AssistantKit to generate platform-specific agent files from the specs.

Install AssistantKit

go install github.com/agentplexus/assistantkit/cmd/assistantkit@latest

Generate Agents

From the repository root:

# Generate using defaults (--target=local, --output=.)
assistantkit generate agents

# Or explicitly specify options
assistantkit generate agents --specs=specs --target=local --output=.

This generates:

• .claude/agents/*.md - Claude Code agent markdown files
• plugins/kiro/agents/*.json - Kiro CLI agent JSON files

CLI Flags

Flag	Default	Description
--specs	specs	Path to specs directory containing agents/ and deployments/
--target	local	Deployment target (looks for specs/deployments/<target>.json)
--output	.	Output base directory (repo root)

Runtime CLI

The content CLI runs the agent team to generate content from conversations.

Build

go build -o content ./cmd/content

Usage

# Generate all content types from a conversation
./content generate --input=conversation.json --output=./output

# Generate specific formats only
./content generate --input=conversation.json --agents=blog,twitter

# List available agents
./content list-agents

Input Format

The input can be a JSON conversation file:

{
  "messages": [
    {"role": "user", "content": "Let's discuss AI agents..."},
    {"role": "assistant", "content": "AI agents are..."}
  ]
}

Or a Markdown file with the conversation.

License

MIT License