What changed in SF Analyzer.
These entries are rendered from the same shared changelog used by the desktop app, so public release notes and the in-app release modal stay aligned.
Implemented SF CLI compatibility guard (Fase 21) with preflight checks, version checks, and warning popups on invalid CLI configurations.
Added registry-safe metadata retrieval to skip unsupported types using the Salesforce source-deploy-retrieve Node registry.
Added the Document Agent Harness (Fase 13) to guide AI agents in drafting TDDs and impact reports using guarded read-only tools, SOQL guards, and coverage guards.
Implemented incremental retrieve logic based on lastModifiedDate (Fase 11) to download only changes, reducing runtimes.
Added resilient analysis runs (Fase 12) featuring exponential backoff retries for transient errors, runtime cancellation, and partial-failure reporting.
Redesigned the Wiki Browser as an Obsidian-lite MD Viewer (Fase 20b) with folders, TOC scroll-spy, wikilinks, and syntax-highlighted code blocks.
Added deep-link sf-analyzer:// protocol support, OS-backed secret keychain storage, and portal limit telemetry.
Added Custom-Focus filtering to exclude standard/managed metadata and isolate custom changes.
Parallelized describeObjects metadata retrieval, reducing analysis time from minutes to seconds.
Implemented local CLI caching for metadata catalog and object describes to make subsequent runs nearly instantaneous.
Added timing logs and step-by-step progress tracking for analysis runs.
Rebuilt the analyzer as a local-first desktop application using Electron and React.
Added project-based local wiki storage following the Karpathy documentation pattern.
Integrated local SQLite database for Salesforce org credentials and cache.
Added interactive Entity-Relationship Diagram (ERD) using ReactFlow to visualize custom object schemas.
Implemented offline-first metadata retrieve and git-backed history tags.
Added Better Auth integration for secure user login and limit tracking.