Om MCP Example Prompts

Practical prompts for Diligence, Data Access, Jobs, Artifacts, and Hub workflows inside Codex or Claude Code.

The best way to use Om MCP is to ask for outcomes, not raw transport details. These examples are written the way a customer or operator would naturally ask for the work.

Diligence prompts

Target triagetext
Use Om MCP to search KRAS and summarize the therapeutic strategy, current
clinical context, and the highest-signal citations.
Deep diligencetext
Run deep diligence on KEAP1 resistance mechanisms and give me the
strongest evidence with citations.

Data Access prompts

Protein discoverytext
Show me which proteins are available in Om Data Access and which ones are
most relevant to EGFR-family work.
Molecular binderstext
Give me molecular binders for this protein UUID and tell me what
downstream workflow you would run next.
Binder plus non-binder comparisontext
For this protein UUID, get molecular binders and non-binders and explain
how you would use them in a ranking or ML workflow.

What Data Access returns

These prompts are written in customer language. Under the hood, Om MCP currently satisfies them through Data Access tools that return signed binder and non-binder dataset export URLs for the requested protein UUID.

Jobs and Hub prompts

Launch and waittext
Launch the appropriate Hub workflow for this target, wait for completion,
and return the result plus the artifact URLs.
Export job resulttext
Export this completed Om job as JSON and summarize the most important
fields for a scientist reviewing the run.

Artifact prompts

Upload an input filetext
Upload this file to Om artifacts and then use it as the input to the next
Hub workflow.

Good prompt patterns

  • Ask for the result you want, not the raw API endpoint.
  • Name the target, protein UUID, or workflow clearly.
  • Ask the client to explain the next step after it retrieves data or launches work.
  • Use follow-up prompts to turn evidence into execution while staying in the same session.