Skip to content

Experimental

This section contains experimental features, tools, and methodologies that are under active development. These are promising approaches that haven’t yet been fully validated or integrated into the core framework.

A comprehensive system for quantifying delegation risk using probability distributions rather than point estimates. Includes:

  • Interactive Tools: Monte Carlo simulators, Bayesian trust updaters, sensitivity analyzers
  • Estimates Registry: Calibrated priors for failure probabilities, damage distributions, and mitigation effectiveness
  • Methodological Guides: Expert elicitation protocols, incident database integration

These features are:

  • Functional but may have rough edges
  • Subject to change based on feedback and testing
  • Not yet validated against real-world deployments

We encourage experimentation and welcome feedback on what works and what doesn’t.

If you’re interested in helping develop these experimental features:

  1. Try them out on your own risk assessments
  2. Report issues or suggestions via GitHub
  3. Share calibration data if you have it
  4. Propose improvements or new tools