hcls-build-agent
Skill277 starsChanged 6 months agoMIT-0View on GitHub
Use when a developer wants to build a new healthcare or life sciences agent, structure tools and system prompts for an HCLS workflow, or create a Strands agent with domain-specific capabilities. Also use when someone asks about agent architecture, tool design, or system prompt patterns for clinical, genomics, or drug discovery use cases.
In this repository9 other files it gives its agents.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 92
When building infrastructure for the agent, use:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Building an HCLS Agent
When to use this skill
- Developer asks "how do I create a new HCLS agent?"
- Developer needs to structure tools, prompts, and workflows for a healthcare domain
- Developer is building an agent for genomics, drug discovery, clinical trials, or other HCLS workflows
Agent Architecture
An HCLS agent is composed of:
Agent = System Prompt + Tools (MCP) + Skills (Knowledge) + GuardrailsSteps
1. Choose a template
| Template | When to use |
agentcore_template/ | Backend-focused: agent runtime + Gateway tools + Streamlit UI |
| FAST | Full-stack: React frontend + Cognito auth + CDK deployment |
2. Define your agent's domain scope
- What HCLS workflow does it address?
- What data sources does it need? (databases, ontologies, literature)
- What actions should it perform? (query, analyze, generate, validate)
- What guardrails are needed? (PHI handling, clinical disclaimers, data validation)
3. Create tools
Tools are Python functions exposed via AgentCore Gateway (Lambda targets) or as local Strands tools.
# Local Strands tool
from strands import tool
@tool
def search_variants(gene: str, significance: str = "pathogenic") -> dict:
"""Search for genetic variants by gene name and clinical significance."""
# Implementation
passFor Gateway tools (accessible to any MCP client), create Lambda functions and register as Gateway targets. See agents_catalog/28-Research-agent-biomni-gateway-tools/ for the pattern.
4. Write the system prompt
Include:
- Domain expertise and role definition
- Available tools and when to use each
- Output format expectations
- Clinical/scientific disclaimers
- Guardrails (what NOT to do)
5. Add MCP server connections
Reference existing MCP servers for domain tools the agent needs:
- Biomedical databases: deploy Biomni Gateway (
mcp-servers/agentcore-gateway/biomni-research-tools/) - Ontology lookup: deploy OLS server (
mcp-servers/agentcore-runtime/ontology-lookup-service/) - Literature: configure PubMed (
mcp-servers/third-party/pubmed/) - Genomics workflows: configure HealthOmics (
mcp-servers/aws-public/aws-healthomics/)
6. Test
# Local testing with Strands
python main.py --prompt "Your test query"
# Test Gateway tools independently
python tests/test_gateway.py --prompt "Your test query"References
- Reference implementation (simple):
agents_catalog/24-Deep-Research-agent/ - Reference implementation (full Gateway):
agents_catalog/28-Research-agent-biomni-gateway-tools/ - Reference implementation (FAST template):
agents_catalog/35-Terminology-agent/ - Strands Agents docs: use the
strands-docsMCP server - AgentCore docs: use the
agentcore-docsMCP server
AWS MCP Servers Used
When building infrastructure for the agent, use:
aws-mcp— create IAM roles, Lambda functions, S3 bucketsagentcore-docs— AgentCore API reference for Gateway/Runtime/Memory configurationaws-healthomics— if the agent needs genomics workflow capabilities
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

