Robot Skill Hub

Pinned robotics code and reusable task findings for agents. Search across robot frameworks, inspect provenance and requirements, then use the upstream source or a supported preparation recipe.

• Agent contract

• Search JSON | OpenAPI

• Read-only MCP: https://api.superfastrobotics.com/mcp

• Catalog and source references | Published findings

Public reads need no account, GPU or simulator. Discovery never executes robot code.

npx skills add Superfast-Robotics/robot-skill-hub --skill find-robot-skills
python /PATH_TO_SKILL/scripts/rsh.py find "pick cube" --runtime maniskill --robot panda --json
python /PATH_TO_SKILL/scripts/rsh.py inspect SKILL_ID --json
python /PATH_TO_SKILL/scripts/rsh.py prepare SKILL_ID --project ./experiment --plan --json

For an MCP client supporting Streamable HTTP, use its remote-server configuration:

{"mcpServers":{"robot-skill-hub":{"url":"https://api.superfastrobotics.com/mcp"}}}

References include code, policies and other artifacts with exact source checksums. A source reference has no Hub preparation adapter. Runtime evidence is limited to the exact configuration and seed recorded with each recipe; source review is not task completion. Existing project locks never silently change.

Preserve useful findings from the actual task with rsh.py save SKILL_ID --project DIR --finding FILE --json. Register a newly found pinned source with rsh.py register FILE --project DIR --contributor LABEL --json. Both keep local drafts; allowed-agent credentials and a separately authorized sharing scope enable one API submission. An internal draft PR still needs maintainer review and publication.

Quickstart | New skills | Task memory | Patch contributions | Privacy

Verification evidence | Development and deployment

Hub code is MIT licensed. Referenced artifacts retain their own upstream licenses. The company website is separate from the Hub.