Amass logo
Build AssistantDocumentationAPI ReferenceApp gallery
Sign inSign up

Search documentation

Search the documentation by keyword

Getting Started
  • Overview
  • Quickstart
  • Authentication
  • Roadmap
  • Pricing
For AI Agents
  • LLM Quick Reference
Amass Cores
  • BiomedCore
  • DrugCore
  • GeneCore
  • PatentCore
  • RegulatoryCore
  • TrialCore
Examples
  • Starter Agent
  • API Workflows
  • Amass SKILL.md

    Amass SKILL.md

    The Amass SKILL.md is an Agent Skill that teaches an AI agent like Claude, Open Code, Codex, Cursor, and any agent harness that supports the SKILL.md format how to use the Amass *Cores API — which Cores to call for which question, the shape of each endpoint, and the conventions to follow when chaining them.

    Drop it into your agent's skills directory and the model will pick it up automatically whenever a task involves clinical trials, drug development, or biomedical research.

    Install

    Check the repository for the latest instructions, but the simplest way is to use the Vercel Skills CLI. After installing the Skills CLI:

    Shell
    npx skills add amass-technologies/public-skills

    This installs every skill published under skills/ in the repository (currently amass-api) into the location your agent reads from.

    To install manually, copy the individual skill directory into your agent's skills folder — each skill is self-contained, with its own SKILL.md and supporting files describing the Amass APIs and recommended call patterns.

    You will need an AMASS_API_KEY available to the agent — see Authentication for how to create one.

    What it gives the agent

    • Direct knowledge of BiomedCore and TrialCore endpoints, including which parameters matter for which kinds of questions.
    • Worked examples of multi-step API workflows (search → enrich → summarize) the agent can adapt instead of inventing.
    • Conventions for citing Amass results back to the user with stable IDs and links.

    For the underlying API contract the skill is built on, see the LLM Quick Reference. For a fully runnable agent that uses these same APIs without the SKILL.md format, see the Starter Agent.