> "Can the right access boundaries be enforced - who can read, who can edit, which artefact can reach which agent?" ^securability ## Resolution > The property of a knowledge artefact's access being controllable - who can read, who can edit, and which agents (human or machine) it can reach. ^definition Not all knowledge is public; not all readers are trusted equally. Securability makes the right access boundaries enforceable. It covers authentication (who is the reader), authorisation (what they may do), and onward reachability (which agents can be fed which artefacts). The Accessible property in the [[FAIR principles]] subsumes securability for authenticated retrieval. ## In practice ```yaml access: read: [engineering, leadership] edit: [architecture-team] agents: [docs-bot] # may be loaded as context not-agents: [external-llm-bot] ``` A property block on the artefact names the boundaries. The platform enforces them at read time, at edit time, and when an agent asks for context. ## Discovery questions Customer-facing prompts that surface this attribute being missing, grouped by the pain each probes. ### From [[Access boundaries fail]] (C) ![[Access boundaries fail#^discovery-questions]]