sim_spec — Claude Code prompt: what it does, how to install
The sim_spec Claude Code prompt turns assembled trial evidence into a parameter sheet for a Monte Carlo model. A separate builder implements that sheet.
no license ★ 654 updated 2026-11-09
/plugin marketplace add anthropics/life-sciences
/plugin install clinical-trial-forecast@life-sciencesWhat sim_spec does
sim_spec converts evidence about a drug and its earlier trials into a parameter sheet for a Monte Carlo model. The sheet covers trial design, parameter distributions, intervals, rationales, and citations. It also checks whether an analogue already includes an effect, so the same risk is not added twice.
The prompt reads files from intake/, knowledge/, topics/, analyses/, and research/ in a prescribed order. It has no web access and does not run code; a separate builder implements the resulting sheet. The command is installed from the life-sciences marketplace as clinical-trial-forecast.
What it helps with
- draft a parameter sheet for a clinical-trial model
- estimate how a phase 2 effect translates to phase 3
- assign uncertainties to parameters without double counting
- record the evidence and rationale behind model parameters
Example prompts for the agent
- Create a parameter sheet for a Monte Carlo model of this clinical trial.
- Review the earlier trial evidence and estimate how the phase 2 effect translates to phase 3.
- Specify the model and show which risks are already covered by its parameters.
How to install sim_spec
1Plugin marketplace
/plugin marketplace add anthropics/life-sciences/plugin install clinical-trial-forecast@life-sciences2Manually: put the files in place
.claude/commands/sim_spec.mdHow to use
Paste the text at the start of a chat or into project instructions
Questions about sim_spec
How do I install sim_spec in Claude Code?
Run /plugin marketplace add anthropics/life-sciences, then /plugin install clinical-trial-forecast@life-sciences. The command file is .claude/commands/sim_spec.md.
What does sim_spec do?
It prepares a parameter sheet for a Monte Carlo clinical-trial model, covering the trial design, distributions, rationales, and citations. It checks that an effect already reflected in an analogue is not counted again.
Can sim_spec browse the web or run code?
No. It reads the assembled evidence and writes a specification for a separate builder to implement.