Tagged: ai-coworker
Preview an AI assistant’s resource scope before you authorize it
Before approving an AI assistant, inspect exactly which workspace resources its credential can read, change, create, and disclose across private and public spaces.
Resource scoped API keys keep AI Coworker writes on target
See how resource scoped API keys limit AI Coworker to the intended project, wiki space, or request board when ambiguous names could send a write elsewhere.
When AI Coworker acts by default and when it asks first
Learn how Dutify decides when AI Coworker should act by default and when a human must confirm, based on reversibility, exposure, and the cost of a wrong write.
Why AI actions need a reproducible decision trail
Learn why reproducible AI actions require records of the original instruction, retrieved passages, model version, and human approval before any task or wiki change.