A Successful JENNER Measles MCP experiment - laser-base/laser-measles GitHub Wiki
Environment
- jenner-measles-mcp running locally (v.0.9.2)
- jenner-mcp running locally (v1.0.?)
- claude code
Prompt
"Using python3.11, and the /var/opt/idm/venv_measles_docs venv, and the MCP servers, and not using any local documentation or spellunking or inspecting, create a series of laser-measles models of increasing complexity and interest, checkpointing each working model with beautiful outputs, proceeding all the way to a sensitivity analysis and maybe even a scientifically interesting result. Use the features advertised in the docs and use model types suitable to the analysis."
(Note that the first part is just to save some on setup; could probably leverage claude.md more.)
Results
Code Artifacts
The five AI-generated scripts from this experiment have been preserved in the laser-mcp repository under:
experiments/jenner_measles_mcp_2026_12_20/
They are pinned to commit:
f048fe7aef5fcd4dc4cfe8d4226cd66fadfb4959
These files represent the exact outputs produced during the MCP experiment and are retained as research artifacts. They are not maintained production examples and may depend on the laser-measles API v. 0.9.0 (as of February 2026).
Checkpoints
1. checkpoint_01_single_patch.py
Baseline single-patch SEIR model. Establishes basic transmission dynamics and visualization output.
2. checkpoint_02_spatial_gravity.py
Introduces multi-patch spatial structure with gravity mixing. Demonstrates spatial heterogeneity and cross-node transmission.
3. checkpoint_03_vital_dynamics.py
Adds births, aging, and maternal immunity. Explores demographic turnover and long-term endemic structure.
4. checkpoint_04_sia_campaigns.py
Implements SIA campaigns and vaccination dynamics. Demonstrates discrete intervention effects and transient susceptibility suppression.
5. checkpoint_05_sensitivity_analysis.py
Performs a structured sensitivity analysis across key parameters. Produces comparative outputs suitable for scientific interpretation.