Choose theme
AI Visuals
LabClaw
LabClaw is a library of scientific and biomedical skills that teach an OpenClaw-compatible agent how to use particular research tools.
The repository describes individual SKILL.md files for biology, data analysis, literature search, visualization and related workflows. It is a skill library, not a complete lab application. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
Scientific instructions for agents
Each skill describes when a tool applies, how to call it and what output to produce.
Why it stands out
Choose skills for a research task
Domain folders let a team select literature, analysis or visualization instructions instead of installing the whole collection.
Availability
Repository with individual skill files
The full collection or selected folders can be added to a compatible agent setup. Tool dependencies remain specific to each skill.
Why it matters
What makes it useful
A researcher planning a single-cell analysis can find the repository's Scanpy and AnnData skills, while a literature workflow can start with its search and citation skills. The folders provide a task-specific starting point for an existing agent.
What to know
Where it fits
The skills sit between a compatible agent and scientific tools or services. Installing the instructions is separate from supplying the runtime, data, software dependencies and credentials those tools require.
Notable points
What stands out
The Scanpy skill documents a quality-control command with .h5ad input and filtered .h5ad output. Its mitochondrial, minimum-gene and minimum-cell thresholds can be recorded alongside the output instead of disappearing inside an agent conversation. Confirm that the referenced script is available in your checkout before using the command.
Before using
What to review
Read the selected skill's requirements rather than assuming the library has one shared setup. For example, PubMed Search uses Valyu's search API and needs a Valyu key plus Node.js; a public database name does not mean the skill calls that database directly.
Check data access and external service calls for that specific workflow. The scientific-diagram skill sends prompts to a Gemini image model through the yunwu.ai relay and needs an API key; those requests are not kept entirely within your local agent.
Scientific and biomedical results still need appropriate expert checks; the library is not a clinical decision service.
Reader fit
Who may find it relevant
Researchers and developers who already have a compatible agent and want reusable instructions for named scientific tasks.
Teams need the domain knowledge to assess procedures and outputs; the library does not replace that expertise.
Editorial note
Why LifeHubber lists it
A scientific figure needs more than an attractive image. LabClaw's diagram skill specifies exact labels, compartments and relationships such as activation and inhibition. Those domain-specific instructions make the library worth a look for researchers turning a mechanism into a draft illustration: they provide a starting specification for what the figure must communicate. The generated biology and labels still need checking.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
What to explore next
Compare skill instructions with a complete research workspace.
LabClaw adds procedures to an existing agent. A local scientific workspace is another route when you want chats, code, figures and a notebook together.
More in Embodied / Physical AI
Keep browsing this category
Explore more embodied AI resources.
LeRobot
huggingface/lerobot
A Hugging Face robotics library for robot-learning workflows, with models, datasets, robot interfaces, training, evaluation, deployment, and v0.6.0 updates across world models, reward models, benchmarks, dataset tooling, and cloud training paths.
LeRobot Humanoid
Virgileboat/lerobot-humanoid
A public humanoid robotics workspace around a low-cost 3D-printed robot-learning platform, with hardware files, assembly documentation, runtime tools, simulation/model assets, identification tooling, and training environments.
elrobot
norma-core/hardware/elrobot
A low-cost 3D-printed robotic arm intended for physical AI research and imitation learning.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
