biomni

name	biomni
description	Use when working from the local Biomni repository to run agent-style biomedical tasks or inspect Biomni's biomedical tool modules and examples.
allowed-tools	Bash, Python

biomni

Workspace-local entry point for the Biomni repository at /home/vimalinx/Projects/bio_studio/repositories/active/Biomni. The repo is present, the top-level Python package can be imported by adding the repo to sys.path, and the README documents both agent-style usage (A1) and a large biomedical tool tree under biomni/tool/.

Quick Start

• Repository: /home/vimalinx/Projects/bio_studio/repositories/active/Biomni
• Primary workflow: activate biomni_e1, configure API keys, then use from biomni.agent import A1
• Fast local check: python -c "import sys; sys.path.insert(0, '/home/vimalinx/Projects/bio_studio/repositories/active/Biomni'); import biomni; print(biomni.__file__)"

When To Use This Tool

• Running Biomni as a biomedical agent from the local repo
• Inspecting Biomni's tool modules for genomics, molecular biology, protocols, literature, and related domains
• Following the repo's own quick-start, profile-switching, or Gradio demo flows
• Auditing whether the current workspace is actually ready for Biomni execution

Common Patterns

# Enter the local repository
cd /home/vimalinx/Projects/bio_studio/repositories/active/Biomni

# Minimal repo-based import check
import sys
sys.path.insert(0, "/home/vimalinx/Projects/bio_studio/repositories/active/Biomni")
import biomni
print(biomni.__file__)

# Agent-style usage from the README
from biomni.agent import A1
agent = A1(path="./data", llm="claude-sonnet-4-20250514")
agent.go("Perform scRNA-seq annotation at [PATH] and generate meaningful hypothesis")

Recommended Workflow

1. Start in the local Biomni repo and follow its environment setup rather than assuming the current bio env is enough.
2. Configure API keys through .env or shell environment variables before trying agent mode.
3. Use the A1 examples in the README for full agent workflows, or inspect biomni/tool/ if you only need module-level capabilities.
4. If you want a GUI, follow the documented Gradio path only after the core Python environment is healthy.

Guardrails

• The repo is present and import biomni works when the repo path is added to sys.path, but deeper tool imports failed locally because langchain_core is missing.
• The README expects a dedicated biomni_e1 conda environment plus additional package installation; do not assume the current workspace Python is sufficient.
• Agent mode requires API-key configuration. The old autogenerated skill text incorrectly implied zero-key library use was generally turnkey.
• The README says the Biomni datalake can auto-download on first agent creation and may require roughly 11GB.
• The Gradio interface is documented as requiring gradio>=5,<6.