# capsid-bench: what it is made of and how it connects | Dustin Edwards

> The study harness that compares Claude Code with and without Capsid on the same tasks, scored on hidden tests.

Source: https://dustinedwards.info/systems/capsid-bench

The study harness that compares Claude Code with and without Capsid on the same tasks, scored on hidden tests.

[github.com/DrDustinEdwards/capsid-bench](https://github.com/DrDustinEdwards/capsid-bench) · [capsid.dustinedwards.info/s/capsid-bench](https://capsid.dustinedwards.info/s/capsid-bench/)

How capsid-bench connects

## Used by 1

- [Capsid](https://dustinedwards.info/systems/capsid) coordinates

## Inside capsid-bench

### Runs

- #### bench-run workflow

  GitHub Actions

  bench-run workflow Runs one task in one arm, behind an egress allowlist, then scores and seals it. writes to MCP Worker (Capsid) reads Claude Code writes to drand and Sigstore

- #### bench-walk workflow

  GitHub Actions

  bench-walk workflow Draws the accepted task set from the frozen frame with a drand seed. reads drand and Sigstore

### Outside

- #### Claude Code

  Claude Code on a subscription token

  Claude Code The agent under study, in both arms.

- #### drand and Sigstore

  drand quicknet, Sigstore and Rekor

  drand and Sigstore The public randomness for the draw, and the time-locked, logged seal.

## Uses 1

- [Capsid](https://dustinedwards.info/systems/capsid) writes to

Public model, 2026-10-23. [Markdown](https://dustinedwards.info/systems/capsid-bench.md)