Skip to content

Add measured Skillspace improvement and grip-retention experiments - #21

Draft
Tarzelf wants to merge 5 commits into
feat/live-piper-workstation-v01from
research/retention-baseline-20261016
Draft

Tarzelf wants to merge 5 commits into
feat/live-piper-workstation-v01from
research/retention-baseline-20261016

Conversation

@Tarzelf

@Tarzelf Tarzelf commented Oct 18, 2026

Copy link
Copy Markdown
Contributor

The simulated arm previously lacked a measured improvement loop and enough evidence to explain the remaining low-force retention failure. This adds a local Skillspace improvement viewer, frozen comparison protocols, grip/recovery diagnostics, and bounded recovery, smoothing and compliant-transfer experiments with compact evidence.

The retained results keep the original physics, force limits, task scoring and deadline explicit. The student remains at 15/16 after the retention experiments; neither candidate was promoted. This is a simplified rigid-box simulation, separate from PiPER shoe studies and physical hardware.

Validation: 149 focused tests passed across the improvement, grip-diagnostic, recovery and retention modules. Twelve loopback HTTP tests passed after rerunning with local-server permission. All 98 changed file versions across the five commits were inspected: text-only, synthetic dataset metadata, no raw footage or detected credentials. Browser UI checks were not rerun for this publication.

This draft publishes the existing five-commit research foundation separately so the native contact audit can be reviewed as a focused dependent change. It is based on the live-workstation branch; it does not merge the earlier simulation stack.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant