Skills 2026-09-27 17:22 UTC
BetterAIChatv0.27.4
Version [v0.27.4] committed by Verlintas. Update Category: Feature Addition & Enhancement. Commit details: docs: add ch27 Design philosophy & development approach (zh+en). This update introduces new capabilities and module extensions.
CommitRaw commit message
docs: add ch27 Design philosophy & development approach (zh+en) New chapter capturing the thinking behind the code for handover: - 8 design principles actually followed (local-first, never trust the model, budgeted text results, actionable failures, deny by default, react-don't-poll, testable seams, single source of truth) - Decision log with trade-offs: hand-rolled DI, built-in tools vs MCP, hand-written SSE parser, Flow-based engine loop, explicit Room migrations, custom tables + pinned markdown lib, self-heal + breaker, layout-driven scrolling, manual signing - The 7-step feature workflow this project actually uses (define observable behavior -> mock scenario first -> implement -> device verify -> targeted tests -> periodic audits -> green CI) - Anti-pattern list with the real incident behind each (main-thread decode, eval-style math, timestamp IDs, isScrollInProgress trap, swallowed cancellation, unbounded loops...) - Closing note on the project's spirit for the next maintainer