v0.83.8
August 21, 2026
Sensitive data protected across all versions, transparent AI Summary. Sensitive-field classification applies to a process's entire version family, history shows what was actually protected, the AI Summary communicates clear states, and integrations get new tools with explicit risk labels.
-
Security
Sensitive fields protected across the whole version family
Classifying a field as sensitive now protects every element of the process, including those still on earlier versions: protection follows the family, not the single version. It applies everywhere data appears: interface, exports, integrations, extension, and AI features.
-
Improvement
Transparent history on applied protection
Each history row states how many fields were actually protected: you can tell an untouched entry from a redacted one at a glance, and the protection date reflects a real change.
-
Improvement
AI Summary with clear states
An element's AI Summary always communicates its real state: preparing, ready, failed (with an understandable reason), disabled, or not requested. An interrupted generation no longer stays "in progress" forever: it recovers on its own.
-
Integration
New MCP tools with risk labels
Clients connected via MCP get three new tools — product version, organization profile, and process analytics (respecting the profile's permissions) — and every tool in the catalog now explicitly declares whether it is read-only, destructive, or outward-facing: the assistant can assess the risk before acting.