Skip to main content

outo-agentcore

outo-agentcore

CLI agent tool powered by agentouto SDK

PyPIQuick StartCLIDevelopment

Features

  • Powered by agentouto: Full peer-to-peer multi-agent SDK with parallel execution, streaming, background agents
  • Agent support: Agents call each other recursively with no orchestrator
  • Skills support: Agent Skills specification compliant - reusable instruction packages
  • Multi-provider: OpenAI, Anthropic, Google Gemini, and any OpenAI-compatible API (Ollama, vLLM, LM Studio, etc.)
  • Session persistence: Continue conversations across runs
  • Bash execution: Built-in shell command tool
  • Wiki knowledge base: Optional OutoWiki integration for persistent knowledge management
  • Simple CLI: No complex TUI, just commands

Installation

# pip
pip install outo-agentcore

# uv (recommended)
uv tool install outo-agentcore

Uninstall

# pip
pip uninstall outo-agentcore

# uv
uv tool uninstall outo-agentcore

To remove all data:

rm -rf ~/.outoac

Quick Start

1. Setup

outoac setup \
  --base-url http://localhost:11434/v1 \
  --api-key your-key \
  --default-model llama4:scout \
  --agent-md ~/.outoac/agents/main.md \
  --default-agent main

2. Create agent definitions

mkdir -p ~/.outoac/agents

Create ~/.outoac/agents/main.md:

# Coordinator

You coordinate work. Call other agents when needed.

Create ~/.outoac/agents/researcher.md:

# Researcher

You find and organize information from the web.

Create ~/.outoac/agents/writer.md:

# Writer

You create well-structured reports from research.

3. Start chatting

outoac chat "What is the capital of France?"

4. Continue a session

outoac sessions                    # List sessions
outoac chat "Tell me more" --session <session-id>

CLI Commands

Command Description
outoac setup Configure providers and agents
outoac chat "message" Start a chat session
outoac chat "message" --session <id> Continue existing session
outoac chat "message" --agent <name> Use specific agent
outoac sessions List all sessions

Global Options

Argument Description Default Example
--config, -c Path to config.json ~/.outoac/config.json ./my-config.json

Setup Options

Argument Description Default Example
--base-url Provider API base URL - http://localhost:11434/v1
--api-key API key for provider - sk-xxx
--default-model Default model for provider - llama4:scout
--provider-name Provider name default openai
--agent-md Path to main agent markdown - ~/.outoac/agents/main.md
--default-agent Default agent for chat main researcher

Chat Options

Argument Description Default Example
message Message to send (required) - "Hello"
--session, -s Session ID to continue - a1b2c3d4...
--agent, -a Agent name to use main researcher
--debug Enable debug output false -

Sessions Options

Argument Description Default Example
--limit Max sessions to show 10 20

Configuration

Default config file: ~/.outoac/config.json

Override with --config flag or OUTOAC_CONFIG environment variable:

# Use custom config
outoac --config ./project-config.json chat "Hello"

# Or via environment variable
export OUTOAC_CONFIG=./project-config.json
outoac chat "Hello"
{
  "providers": {
    "default": {
      "kind": "openai",
      "base_url": "http://localhost:11434/v1",
      "api_key": "your-key",
      "default_model": "llama4:scout",
      "max_output_tokens": 0
    }
  },
  "agents": {
    "main": "~/.outoac/agents/main.md",
    "researcher": "~/.outoac/agents/researcher.md",
    "writer": "~/.outoac/agents/writer.md",
    "reviewer": "~/.outoac/agents/reviewer.md",
    "coder": "~/.outoac/agents/coder.md",
    "tester": "~/.outoac/agents/tester.md"
  },
  "default_agent": "main",
  "skills_dir": "~/.outoac/skills/",
  "wiki": {
    "enabled": false,
    "wiki_path": "~/.outoac/wiki/",
    "provider": "openai",
    "model": "gpt-5.5",
    "api_key": "",
    "base_url": "",
    "max_output_tokens": 0,
    "debug": false
  }
}

Provider Settings

Parameter Description Default
kind Provider type (openai, anthropic) openai
base_url API endpoint URL -
api_key API key for authentication -
default_model Default model for agents -
max_output_tokens Default max output tokens (auto-detected if 0) 0

Note: If max_output_tokens is 0 or not set, the system automatically retrieves the optimal value from the LCW API. Agent-level settings override provider defaults.

Wiki Settings

Parameter Description Default
enabled Enable wiki knowledge base tools false
wiki_path Path to wiki directory ~/.outoac/wiki/
provider LLM provider for wiki operations openai
model Model for wiki analysis (empty, uses provider default)
api_key API key for wiki provider (empty, uses provider key)
base_url Base URL for wiki provider (empty, uses provider URL)
max_output_tokens Max tokens for wiki responses (auto-detected if not set) 0 (auto)
debug Enable wiki debug logging false

Note: If max_output_tokens is 0 or not set, the system automatically retrieves the optimal value from the LCW API.

Agent Markdown Format

Agent definitions use markdown with optional YAML frontmatter:

---
model: MiniMax-M2.7
provider: default
temperature: 1
max_output_tokens: 4000
---

# Research Agent

You are a research specialist. Your job is to find information
and provide detailed analysis.

Frontmatter fields:

  • model: Model name (overrides provider default)
  • provider: Provider name to use
  • temperature: Sampling temperature (0.0-2.0)
  • max_output_tokens: Maximum output tokens (auto-detected if not set)

Body:

  • First # heading becomes the agent's role
  • Rest becomes the agent's instructions

Note: If max_output_tokens is not set, the system automatically retrieves the optimal value from the LCW API.

Agent Example

Create multiple agents:

~/.outoac/agents/main.md:

# Coordinator

You coordinate work. Call other agents when needed.

~/.outoac/agents/researcher.md:

# Researcher

You find and organize information from the web.

~/.outoac/agents/writer.md:

# Writer

You create well-structured reports from research.

Setup with all agents:

outoac setup \
  --base-url http://localhost:11434/v1 \
  --api-key your-key \
  --default-model llama4:scout \
  --agent-md ~/.outoac/agents/main.md \
  --default-agent main

Now main agent can call researcher and writer using call_agent tool.

Supported Providers

Provider Base URL Default Model
OpenAI https://api.openai.com/v1 gpt-5.5
Anthropic https://api.anthropic.com/v1 claude-sonnet-4-6
Google Gemini https://generativelanguage.googleapis.com/v1beta gemini-3-1-pro
Ollama http://localhost:11434/v1 llama4:scout
LM Studio http://localhost:1234/v1 (local model)
vLLM http://localhost:8000/v1 (local model)

Wiki Integration

Outo-agentcore integrates with OutoWiki for persistent knowledge management. When enabled, agents can record and search information across conversations.

Enable Wiki

Add to ~/.outoac/config.json:

{
  "wiki": {
    "enabled": true,
    "provider": "openai",
    "model": "gpt-5.5",
    "api_key": "sk-..."
  }
}

Note: If the wiki section is not present in the config, wiki features are disabled by default.

Wiki Tools

When enabled, agents have access to:

  • wiki_record: Save information to the wiki for future reference
  • wiki_search: Search the wiki for relevant knowledge

Use Cases

  • Remember user preferences across sessions
  • Build a knowledge base from conversations
  • Search for previously discussed topics
  • Maintain context for long-running projects

Skills

Skills are reusable instruction packages that extend agent capabilities. They follow the Agent Skills specification.

Creating Skills

mkdir -p ~/.outoac/skills/my-skill

Create ~/.outoac/skills/my-skill/SKILL.md:

---
name: my-skill
description: What this skill does and when to use it.
---

# Instructions

Your detailed instructions go here...

How Skills Work

  • Skills are automatically discovered from ~/.outoac/skills/
  • Agents see all available skills in their system prompt
  • Agents automatically activate skills based on task description
  • Skills are loaded progressively (metadata → instructions → resources)

See Skills Documentation for details.

Built-in Tools

Tool Description
bash Execute shell commands
call_agent Call another agent
finish Return final result
wiki_record Record information to wiki (when enabled)
wiki_search Search wiki knowledge base (when enabled)

Directory Structure

~/.outoac/
├── config.json          # Provider and agent settings
├── agents/              # Agent markdown definitions
│   ├── main.md
│   ├── researcher.md
│   ├── writer.md
│   ├── reviewer.md
│   ├── coder.md
│   └── tester.md
├── skills/              # Skill definitions (Agent Skills spec)
│   ├── code-review/
│   │   └── SKILL.md
│   └── git-workflow/
│       └── SKILL.md
├── sessions/            # Session persistence
│   └── <session-id>.json
└── wiki/                # Wiki knowledge base (when enabled)
    └── *.md             # Wiki documents

Development

# Clone
git clone <repo-url>
cd outo-agentcore

# Install dependencies
uv sync

# Run tests
uv run pytest tests/ -v

# Run CLI
uv run python -m outoac --help

License

Apache-2.0

Release files for outo-agentcore 0.4.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for outo-agentcore 0.4.0
File Size Uploaded
outo_agentcore-0.4.0.tar.gz 167.2 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for outo-agentcore 0.4.0
File Interpreter ABI Platform
outo_agentcore-0.4.0-py3-none-any.whl Python 3 none any Details

Total release size: 182.5 kB

Release files / outo_agentcore-0.4.0.tar.gz

Download URL outo_agentcore-0.4.0.tar.gz
Size 167.2 kB
Tags Source
SHA-256 checksum
How to use checksums
443bac9571961ced5e43b2ab03bac633d2cde16b98cb374931c6deaa0a6807f6
BLAKE2b-256 checksum
How to use checksums
b95126a1228179a1f3a0cf042136637eb707629f97c37a8e4c752033e75ff9dc
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on May 10, 2026.

Transparency log

Release files / outo_agentcore-0.4.0-py3-none-any.whl

Download URL outo_agentcore-0.4.0-py3-none-any.whl
Size 15.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
4f9411cd33014b4531af0397ecd3f0c4487924e474ab89c717dd9fbc616fe6b9
BLAKE2b-256 checksum
How to use checksums
dab01b50d19259997a509083e7f63b81e5c4ff0a51e71c825992a9c1f62557de
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on May 10, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.4.0 This release

2 release files

0.3.0

2 release files

0.2.0

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page