AI Stop Before Not Narrate After

Use whenever an agent works under a task-routing policy (a table that says which class of task goes to which model, role, or capability level) and the agent can be handed a task that the table routes elsewhere. Covers the moment a task arrives, the moment the agent notices it is the wrong hand for it, and the boundary between parts of a task. Load when a header, signature, or routing field claims which model or role is doing the work — the agent cannot measure that about itself, and the skill says what it can measure instead.

nowespojrzenie Updated

File contents

nowespojrzenie/HEXAGRAM commit 9e91c75216

Frequently asked questions

npx get-skill add nowespojrzenie/ai-stop-before-not-narrate-after