osborn 0.8.30 → 0.8.31

package/.claude/skills/learned-behaviors/SKILL.md

@@ -4,38 +4,42 @@ Auto-extracted from voice sessions via PreCompact.
 Last updated: 2026-05-01 | Session: unknown...
 
 USER CORRECTIONS:
-- First write attempt for interruption timer was DENIED because it replaced existing LiveKit code instead of adding on top — always ADD fallbacks, never replace existing mechanisms; user said "you're basically deleting what we have"
-- User denied permission dialog for file edit because they wanted to verify the approach first before committing — ask "should I build this?" before delegating write tasks on significant changes
-- When proposing solution approach, verify against LiveKit SDK source AND docs before implementing — user asked "can you verify that we're doing it the right way in terms of live kits"
+- Never delete existing code when adding new functionality — always add as a fallback/layer on top; user denied permission twice when replacements were sent instead of additions
+- Don't dismiss log warnings as "contamination" without verifying — the 0.8.30 reference in logs was real and was incorrectly dismissed
+- Share URLs directly in the message, not just in session files — "send it to me in the actual message"
+- Don't say "I can't read images from URLs in this context" — fetch the public URL and use the Read tool on the saved file
+- When user says "you updated it" trust the session history and confirm rather than expressing surprise
 
 USER PREFERENCES:
-- Voice transcription: "MPM"=npm, "mobile publish"=npm publish, "live gate"=LiveKit, "left kit"=LiveKit, "basketball man"=bash command, "Osborne"=osborn — interpret flexibly
-- Sprite is the publish path for npm going forward (not Mac), set up with automation token
-- Prefers additive code changes: keep existing mechanisms as primary, add new approaches as fallback tiers
-- Wants to understand the "why" before approving implementation — explain the mechanism, not just the solution
-- Wants confirmation before committing to solutions when uncertain: "before we commit to doing this"
+- Prefers additive changes over replacements — keep existing SDK mechanisms as primary, add custom fallbacks only when SDK fails
+- Wants direct URL in voice message (readable), not just written to session files
+- Expects natural chain of: bump version → build → commit → push → publish in one flow
+- Prefers push-based deploy (Railway auto-deploys from GitHub push) over manual steps
+- Wants feedback on async operations — "is it done?" is the recurring question
+- Voice transcription artifacts: "MPM"=npm, "Osborne"=Osborn, "Brazil"=unclear, "basket man"=bash command
 
 DOMAIN KNOWLEDGE:
-- `osborn --version` has no version flag — it starts a new LiveKit-connected instance; use `npm list -g osborn` or `npm list -g osborn --json` for version
-- npm automation token (not web login token) required for headless publish — web login token still requires 2FA for publish
-- `playbackPosition` is 0 because interrupt often fires before TTS audio frames start flowing (`pushedDuration = 0`); also race condition between `addDoneCallback` and `waitForPlayoutTask`
-- `skipTTSQueue: true` does NOT bypass audio frame tracking — `ttsTask` code path is identical
-- `EVENT_PLAYBACK_STARTED` ('playbackStarted') is public, fires on first audio frame, accessible via `audioOutput.on()`
-- OpenAI TTS-1 rate: ~14 chars/second confirmed from community benchmarks
-- `check-version` in sandbox/route.ts was incorrectly using `osborn --version` (starts instance); fixed to `npm list -g osborn --json`
-- execInSprite works any time the Sprite container is running, independent of osborn agent health
-- `checkVersion()` already existed in dashboard/page.tsx (lines 325-338); `installedVersion` and `latestVersion` state already existed
-- Dashboard handleRestart/handleUpdate were not reading API response bodies — server returns rich data but client discarded it
-- Railway auto-deploys from GitHub main branch push
+- `osborn --version` has no version flag — calling it starts a new LiveKit agent instance and connects to a room
+- npm automation token required for headless publish; web login token still requires interactive 2FA
+- `(currentSession as any)?.output?.audio` is the correct path to AudioOutput; `_activity.agentSession.output.audio` is WRONG
+- OpenAI TTS-1 does not provide `synchronizedTranscript` or meaningful `playbackPosition` — both are 0/null
+- `EVENT_PLAYBACK_STARTED` ('playbackStarted') is a public LiveKit AudioOutput event that fires on first audio frame
+- 14 chars/sec is the correct estimate for OpenAI TTS-1 (175-178 WPM × 5 chars/word / 60s)
+- Dashboard `handleRestart`/`handleUpdate` server already does `waitForHealth` before responding; client was ignoring the response body entirely
+- `execInSprite` works any time the Sprite container is running, independent of osborn agent health
+- `npm list -g osborn --json` is the correct way to get installed version (not `osborn --version`)
+- Writer sub-agents will autonomously fix bugs and publish if they have npm auth and the task implies it — need to scope tasks tightly
 
 EFFECTIVE PATTERNS:
-- Layered fallback approach: preserve existing SDK mechanisms, add custom approach only as fallback — explicit user preference confirmed
-- Taking checkpoint before significant changes gives confidence to experiment
-- Reading API response bodies and surfacing errors/success messages is a simple fix for "silent" operations
+- Four-tier fallback chain (sdk-transcript → sdk-position → wall-clock → full-block) preserves future compatibility while adding current fix
+- Checkpoint before significant code changes; user approved checkpoints as a safety measure
+- Delegating "research then build" to researcher first, then writer prevents over-building on incorrect assumptions
+- Reading the actual image file via WebFetch + Read instead of asking user to describe it
 
 ANTI-PATTERNS:
-- Running `osborn --version` on the Sprite created stray processes connected to LiveKit room — never run the osborn binary to check version
-- Replacing existing LiveKit code with custom implementation instead of adding on top — user explicitly denied and corrected this
-- Delegating write task without confirming approach with user first on novel/uncertain changes
-- Not reading the API response body after `await fetch()` — silent failure mode
+- Running `osborn --version` — starts a live agent instance, creates stray processes in LiveKit room
+- Telling user to look for a permission dialog in their browser when the dialog may have already timed out — retry the agent instead
+- Dismissing "contamination" in logs without checking if it's real data — cost a full investigative loop
+- Scoping writer agent tasks too broadly — the dashboard writer autonomously bumped version and published because npm was authenticated and the task was ambiguous about scope
+- Ending a turn with only "opaque work is running" and no conversational engagement
 </summary>

package/dist/index.js

@@ -2994,9 +2994,10 @@ async function main() {
                     }
                     else {
                         try {
-                            const webhookBase = process.env.FLY_APP_NAME
-                                ? `https://${process.env.FLY_APP_NAME}.fly.dev`
-                                : `http://localhost:${apiPort}`;
+                            const webhookBase = data.webhookBase ||
+                                (process.env.FLY_APP_NAME
+                                    ? `https://${process.env.FLY_APP_NAME}.fly.dev`
+                                    : `http://localhost:${apiPort}`);
                             await sendToFrontend({ type: 'meeting_joining', message: 'Osborn is joining your meeting...' });
                             const botId = await recallJoin.joinMeeting(meetingUrl, webhookBase);
                             const sessionId = currentLLM?.sessionId || currentResumeSessionId || 'default';

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "osborn",
-  "version": "0.8.30",
+  "version": "0.8.31",
   "description": "Voice AI coding assistant - local agent that connects to Osborn frontend",
   "type": "module",
   "bin": {