skill
bioos-navigen
A specialized, pluggable system prompts to enhance general AI agent as Bio-OS expert designed to operate on the Bio-OS platforms.
Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/GBA-BI/bioos-navigen itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 26, 2026Bio-OS Navigen Agent
Bio-OS Navigen is a skill-based agent instruction set for operating Bio-OS workflows through natural language. It separates scientific routing from execution details, so the same bioinformatics workflow knowledge can run through either MCP/plugin tools or the pybioos CLI.
Skill Layout
This repository maintains three complementary skill trees:
| Directory | Use Case | Execution Surface |
|---|---|---|
skills-bioinformatics/ | Domain/business-layer skills for genomics, transcriptomics, single-cell analysis, microbiome, pathogen analysis, early warning, proteomics, and Bio-OS image catalogs. | Execution-neutral. These skills route scientific intent and refer to stable skill names such as bioos_platform_operator, without depending on whether the backend is MCP/plugin or CLI. |
skills/ | Platform-operation skills for agents connected to Bio-OS through tools exposed by bioos-mcp-server or the OpenClaw BioOS plugin. | Uses tool names such as submit_workflow, upload_files_to_workspace, get_workflow_logs, and create_iesapp. |
skills-cli/ | Platform-operation skills for agents that call pybioos directly from a shell. | Uses concrete CLI commands such as bioos workflow submit ... and bioos file upload .... This tree includes bioos_cli_locator to verify CLI availability and authentication. |
The intended pairing is:
- MCP/plugin mode: install
skills/plusskills-bioinformatics/. - Direct CLI mode: install
skills-cli/plusskills-bioinformatics/.
Use only one platform-operation tree at a time: choose skills/ if the machine is wired to bioos-mcp-server or the OpenClaw plugin, and choose skills-cli/ if the machine only needs to call pybioos directly. You do not need to install both.
skills/ and skills-cli/ should stay behaviorally synchronized. skills-bioinformatics/ should remain neutral and should not mention either directory by name; that is the small trick that lets the same business-layer workflows work with both execution surfaces.
Installation
To load these skills into your developer assistant, you can use the automatic setup script or configure them manually. The installation is configured at the project level.
Quick Install (Project-Level)
To automatically install these skills into your project directory:
-
Clone or pull the
bioos-navigenrepository to your local machine:git clone https://github.com/GBA-BI/bioos-navigen.git -
Enter your project directory, copy the
setup_skills.shscript, grant execution permissions, and run it by specifying the source repository path:# Enter your project directory cd /path/to/your/project # Copy setup_skills.sh from the cloned repo cp /path/to/bioos-navigen/setup_skills.sh . # Grant execution permission chmod +x setup_skills.sh # Run setup and specify the source path of bioos-navigen ./setup_skills.sh --copy --source /path/to/bioos-navigen
This script will automatically detect and install the skills into the local config folders of the following developers' assistants in your project:
- Antigravity (
.agent/skills/) - Codex (
.agents/skills/and.codex/skills/) - Claude Code (
.claude/skills/)
Manual Installation (Project-Level)
If you prefer to configure the skills manually, run the following commands from the repository root.
Copying Skills Physically (Deployment)
# Create local agent configuration directories
mkdir -p .agent/skills .agents/skills .codex/skills .claude/skills
# Copy skills (e.g. for Antigravity)
cp -R skills-cli/* .agent/skills/
cp -R skills-bioinformatics/* .agent/skills/
Linking Skills Relatively (Development)
To link custom skills without duplicating files, you can create relative symlinks:
# Create directories if not exists
mkdir -p .agent/skills
# Create relative symlinks
cd .agent/skills
ln -s ../../skills-cli/* .
ln -s ../../skills-bioinformatics/* .
Architecture
Bio-OS Navigen uses a three-tier capability model:
| Tier | Layer | Examples |
|---|---|---|
| Tier 1 | Atomic platform capabilities | bioos_data_fetcher, bioos_docker_builder, bioos_wdl_scripter, bioos_workspace_parser, bioos_platform_operator |
| Tier 2 | Pipeline orchestration | bioos_pipeline_developer |
| Tier 3 | End-to-end SOPs | bioos_paper2workspace, bioos_workspace2paper |
The bioinformatics skills sit above this stack: they choose the right workflow and hand execution to the synchronized platform-operation skills.
Related Projects
pybioos: install withpip install pybioos; provides the rootbioosCLI used byskills-cli/.bioos-mcp-server: GBA-BI/bioos-mcp-server; exposes Bio-OS operations as MCP tools used byskills/.bioos-claw: GBA-BI/BioOS-Claws; OpenClaw tools aligned with the MCP tool names.
Development Notes
- Keep workflow-facing file-input behavior synchronized between
skills/andskills-cli/. - Keep business-layer guidance in
skills-bioinformatics/execution-neutral. - In CLI mode, use
bioos_wdl_scripterfor WDL generation, remote validation, and diagnosis of WDL-related run failures. - The legacy modular prompt files live under
system_prompt/, andGEMINI.mdis the compiled single-file prompt for agents that cannot load skills directly.
Contact
For feedback or issues during the trial phase, please contact the development team.