Greenloop
verification-first work around state, drift, evidence, and what it means for an agent to actually be done.
code / research / trust
systems
archive / living
Verification-first tools and research lineages asking whether complex work can stay in contact with its objective.
Greenloop
verification-first workflow framework for AI coding agents. State persists. Claims meet checks. Drift becomes observable. 'Done' has to survive contact with evidence.
Spec tools plan. GREENLOOP proves done.
mcog
metacognition instrument for intent drift, grounding, scope, and completion. A way of watching complex AI-assisted work lose contact with its objective, then measuring whether it can recover.
observation → hypothesis → execution → drift detection → re-groundingThyrsus
the trust layer after reduction. Creation gets something across. Reduction finds contact. Trust is what lets contact become continuation instead of another isolated artifact.
source / reduction / trust / continuation