Runs

Written for version 0.1.14. Records: testing/runs. Every person named in them is a simulated persona.

One line per run, newest first. Each record has a scorecard (what worked, what partly worked, what failed, with the evidence) and run notes (setup, environment, and anything the orchestrator got wrong).

• 2026-10-27, plugin 0.1.9, Targeted re-test (Test B). The 0.1.9 changes: a change review that reads every file, package documentation for the installed version, attribution only with a record, and the empty amendments section at locking, each against a planted defect. scorecard · run notes · setup.
• 2026-10-26, plugin 0.1.8, User simulation (Test B), continued to publication. Two student leads, a novice and an R user, from nothing to a locked plan and a first analysis; then the PI as lead to submission, and a revision lead to publication. scorecard · continuation scorecard · run notes.
• 2026-10-26, plugin 0.1.8, Blind review (Test A). All five reviewers on the simulated project’s repository at commit 16f9124. Results in the 0.1.8 scorecard; full record not in this repository.

Earlier checks without a record here

These are described in VALIDATION.md, but their records are not in this repository:

• Blind review (Test A) of four reviewers (clinician reader, clinician-researcher, statistician, coding expert) on Cryptococcus-cost-analysis at commit 015aa6b.
• Blind review of the plugin itself (0.1.7, 2026-10-26): five reviewers read only the shipped plugin and its stated principles; 46 findings, which 0.1.8 acted on.