v0.51.544 — Release TC (opt-in render markdown in user messages)

Release v0.51.544 — Release TC (opt-in: render markdown in your own messages)

Ships #3933 (rodboev). Maintainer-approved concept (opt-in, default-off). Deep-gated for the two properties that matter: inert-when-off and XSS-safe-when-on.

Added

• Optional setting to render markdown in your own sent messages (#3870). By default, user message bubbles show typed markdown as-is (**bold**, - lists, [links](url) stay literal — code fences and math already render). A new Settings → Conversation → "Render markdown in user messages" toggle (default OFF) routes user text through the same renderMd() sanitizer assistant messages use, so your own markdown renders too. Fully opt-in and self-contained. Thanks @rodboev.

Gate

• Full pytest suite: 9837 passed, 0 failed
• Codex: SAFE TO SHIP — default/off path is byte-identical to master (verified at all 3 sites); ON reuses the existing allowlist sanitizer that rejects javascript:/data:/vbscript: URLs
• Opus: SAFE — SHIP — adversarially drove the real renderMd with XSS payloads (all blocked); both render cache layers correctly invalidated on toggle; meets the "zero chance it breaks anyone who doesn't opt in" bar
• Visual: before/after confirmed — OFF shows raw **markdown**, ON renders bold/list/link cleanly

Closes #3870.