[kinqsradio]

Summary

• features/thread/composer/: Composer, chips, extras, traits, model picker, effort glyph, input/markup/code-block helpers and their tests move in; Composer.tsx goes from 1,109 to 597 lines.
• features/thread/trail/: ToolTrail.tsx goes from 876 to 261 lines; PlanRow.tsx and blocks/ (one file per renderer or small group) split out.
• No visible change: same CSS classes, copy and keys. features/thread/index.ts exports the same names (paths only).
• Baseline: Composer.tsx and ToolTrail.tsx leave the long-files list (13 to 11 long TS files); direction entries stay 21 (paths renamed).

Pure moves: commit 1d1268b only (listed in .git-blame-ignore-revs). Not pure moves:

• Import-path fix commit (relative paths, vi.mock paths, baseline path rename).
• Composer split: verbatim cuts into MentionMenu, composerConstants, suggestions, composerKeys, composerPaste, useAttachments, useStash, useStopRun, useThreadSettings. JSX pieces became components or a function with props: CodeBar, AttachmentStrip, ComposerContext, ComposerField, ComposerBar, controlsMenu. Hook calls moved out of Composer keep their effect order; plain state and selector hooks sit in other places than before. useOwnCommand stays in Composer.tsx so no new app/shortcuts baseline entry appears.
• Trail split: renderers for each call kind became plain functions over a shared BlockContext; Block keeps its hooks. PlanRow now imported from trail/PlanRow.
• Docs: docs/refactor.md status, file paths in docs/design-inventory.md and docs/parity.md (line numbers there are unchanged and now approximate).

Parity rows

none

Frames built

Frames built: none

Testing

• CI covers the rest.
• Local: npm run check, npm test (1714), npm run build, mock Playwright tool-calls, workspace, pane-resize, reading-layout (44 passed). layouts.test.tsx failed once in a full parallel run and passed alone and on rerun.
• Not tested: packaged app, real-engine specs.

Co-authors

none

AI usage

• Written by hand, with no AI
• Harness and version: Claude Code 2.1.318
• Models: claude-sonnet-5-5 (all)
• Effort: medium
• Human role: assigned the refactor slice and its rules; reviews and merges.
• Sources: none

Checklist

• Works on Windows and macOS
• web/src/generated with any contract change; docs/architecture.md with any rule change
• No copied source, secrets, credentials or personal data; no AI attribution in commits, title or branch