Human/agent write lock

A cooperative advisory lock (e.g. /tmp/darkfactory-ontology.lock) that autonomous loops must acquire before rewriting shared brain state, so a human's concurrent hand-edits are not silently clobbered.

grounded in: ontology_learn.py / ontology_maintain.py autonomously rewrite brain/ontology.json on cron; per the project's own edit-lock discipline, hand-edits must hold /tmp/darkfactory-ontology.lock or those loop

Connected concepts

Harness engineering, Guardrails / bounded autonomy, Ontology maintenance, Context ontology, Scheduled autonomy, Human-on-the-loop oversight