An AI agent for therapeutic discovery that executes transparent, multi-step omics analyses including research planning, code execution, and literature reasoning.
Install with the open skills CLI (global, non-interactive — available in every Claude Code session):
npx skills add FreedomIntelligence/OpenClaw-Medical-Skills --skill "medea-therapeutic-discovery" -g -a claude-code -yOr manually — copy the SKILL.md below into:
~/.claude/skills/medea-therapeutic-discovery/SKILL.md---
name: medea-therapeutic-discovery
description: An AI agent for therapeutic discovery that executes transparent, multi-step omics analyses including research planning, code execution, and literature reasoning.
license: MIT
metadata:
author: Artificial Intelligence Group (Adapted from openscientist.ai)
version: "1.0.0"
compatibility:
- system: Python 3.10+
allowed-tools:
- run_shell_command
- read_file
- web_fetch
---
# Medea Therapeutic Discovery Agent
Medea is a multi-stage AI agent designed for therapeutic discovery, modeled after 2026 state-of-the-art open source architectures. It executes transparent, multi-step omics analyses.
## When to Use This Skill
* "Run multi-omics therapeutic discovery pipeline"
* "Analyze omics data for novel drug targets using Medea"
* "Perform literature reasoning and consensus reconciliation for target X"
## Core Capabilities
1. **Research Planning**: Formulates step-by-step omics analysis plans.
2. **Code Execution**: Generates and executes Python/R scripts for data processing.
3. **Literature Reasoning**: Retrieves and synthesizes current literature.
4. **Consensus Stage**: Reconciles experimental evidence with literature to propose high-confidence targets.
## Workflow
1. **Step 1**: Initialize Medea agent with target disease or omics dataset.
2. **Step 2**: Execute the multi-stage pipeline across planning, coding, literature review, and consensus validation.
## Example Usage
**User**: "Run Medea analysis on the provided breast cancer multi-omics dataset."
**Agent Action**:
```bash
python3 -m medea.agent --dataset breast_cancer_omics.h5ad --mode full_discovery
```
Give the agent its own dedicated email inbox via AgentMail. Send, receive, and manage email autonomously using agent-owned email addresses (e.g. hermes-agent@agentmail.to).
Search arXiv papers by keyword, author, category, or ID.
AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.