release 0.1.0-beta.3

v0.1.0-beta.2 was tagged but never published: the release workflow failed on the
windows-x64 build, so no release object and no artifacts exist under that tag.
The fix is in df5f9f8. Burning the version is cheaper than moving a tag that is
already on the remote.

ROADMAP records what happened, and development.md now says plainly that a green
local `bun run release` is not proof — the Windows host takes a different branch
from the Ubuntu runner CI releases from.

ROADMAP.md

@@ -49,10 +49,15 @@ reported rather than fatal.
 **Distribution** — five-platform cross-compiled binaries, checksums, install scripts, CI on
 three operating systems, tag-driven releases.
 
-### 0.1.0-beta.2
+### 0.1.0-beta.3
 
 Fourteen built-in tools, up from six, with sets so the schema cost stays controllable.
 
+`v0.1.0-beta.2` was tagged and never published: `bun build --compile
+--target=bun-windows-x64` rejects `--windows-title` unless the host is Windows, and CI
+releases every target from one Ubuntu runner. It passed locally and failed on the last of
+five builds. The version was burned rather than moving a published tag.
+
 **Visible process** — reasoning streams to a collapsed panel with an estimated token count,
 `ctrl-r` expands it, and it leaves with the turn since it is progress rather than the answer.
 The tool in flight is named from `tool-input-start`, before its arguments have finished