PCVR Experiment Integration

SkillDocs & knowledge

Add TAAC PCVR experiments or change their model, configuration, or checkpoint integration contracts.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the PCVR Experiment Integration skill

What this skill tells your AI

The instructions your AI receives, as published by puiching-memory/taac_2026 in .agents/skills/pcvr-experiment-integration/SKILL.md and read by ahel’s review.

For package wiring, start with experiment integration and the affected package's __init__.py and model.py. For a shared contract change, use architecture to locate its owner. Routine implementation changes within an existing contract do not need this integration workflow.

  • Trace model/config changes through checkpoint reconstruction in evaluation and inference, including the experiment's typed defaults. Successful training alone does not establish that the saved model can be reloaded.
  • Include new experiments in the existing contract-test cases so discovery, model inputs/outputs, and checkpoint reconstruction are exercised.
  • For bundle integration, follow the bundle guide; packaged execution imports the shipped framework and selected experiment.

Select contract tests and the CPU roundtrip from the testing guide. Keep test paths and commands there rather than maintaining a second list in this skill.

Signals

GitHub stars
315
Forks
63
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
pcvr-experiment-integration
Source
github.com/puiching-memory/taac_2026
PCVR Experiment Integration: Skill · ahel