v0.4.35

multica-ai/multicav0.4.35Aug 26, 2026by github-actions[bot]

AI Summary

This release focuses on refining user interface controls and stability, adding conversation management commands and inbox filters, while reverting recent experimental features like HTML preview and Codex retry logic.

Key Highlights

Added `/new` and `/clear` conversation controls
Added From and unread-only filters to inbox
Reverted experimental HTML preview and Codex retry features
Renamed starter prompts to conversation starters

New Features

Per-agent starter prompts for chat
Readable workdir paths in daemon
Conversation starters discoverable from chat
Project display for run_only autopilots
Runtime access enforcement on task claims

Full Release Notes

## Changelog
* e6013e83e466af48afdd3d5c901071d191e01892 MUL-5928: fix(issues): bound board description previews (#6634)
* 8fda4f22d5429e8ece1fc56fdc7ef10cf4fd49d2 MUL-6016: feat(task): retry exact Codex capacity failures (#6747)
* 174c6525899ab43668ce2ead489d73b581643c74 MUL-6141: feat(runtimes): price Qwen/Kimi/Ark models and strip ':' provider prefixes (#6915)
* 0c9f50d1f47a316e75d2accb1255ebeed2a0f818 MUL-6385: feat(chat): add editable live HTML preview (#7188)
* f8ec870f305b4bd3ec2c1be9462dd921426df7a4 MUL-6629: feat(agents): add per-agent starter prompts for chat (#7261)
* 54027ba763fa7da0699b2fe89df4a6b2c13d1c6f MUL-6639: fix(skills): metadata-only skill listings and stall-based timeouts (#7574)
* 9e2d9b6d1ced12d62b4e7fe3e48d10be5caa4e6f MUL-6661 feat(channels): add /new and /clear conversation controls (#7468)
* 6961b6091aa48f7c53c5a67f09d9e25236f2eca7 MUL-6669 feat(inbox): add From and unread-only filters (#7537)
* 6cbb50378f0256dc201741e66a2ca6f92676cb6d MUL-6680: fix autopilot list subscribers + make CLI triggers discoverable (#7552)
* cd169947b5e79d382d786161844485be80940e1b MUL-6686 feat(daemon): use readable workdir paths (#7316)
* bbba9c5460d45803a559ffd7af7238f3628490dc MUL-6697: enforce runtime access on task claims (#7571)
* 26c04803c8d1f92345c528e5d1ebc692ed9aadb1 MUL-6698: feat(agents): make conversation starters discoverable from the chat they run in (#7573)
* 3b1018e9c9ed989004384a480161d132bdb06e46 Revert "MUL-6016: feat(task): retry exact Codex capacity failures (#6747)" (#7586)
* a9d86af46ce90acb5d5a21bebcf33974b0874291 Revert "MUL-6385: feat(chat): add editable live HTML preview (#7188)" (#7584)
* c7d66071e1e7655642f1b431ec51dbc6cc5c534a docs(changelog): add v0.4.35 release entry (2026-08-26) (MUL-6709) (#7583)
* 21b938bfdb88dc502ee923c597a162d4af1415fa docs(changelog): drop reverted entries from v0.4.35 (MUL-6709) (#7589)
* 27610933e2114bf2a1a1c57300cc21a0271275ef fix(autopilots): show and keep Project for run_only autopilots (MUL-6681) (#7553)
* baf1bbf3401b688c82f85230c4d1e3cce6be9554 fix(billing): make Cloud seat enforcement automatic (#7542)
* 6c8f02aeafb0ea23b399f7aa78a76cb3728e4d45 fix(billing): prefer trusted Pro limits over stale quota (#7576)
* 04fd64b410958e2baef1636d49bbc8d9e03c16e7 fix(chat): keep following rapid live output (#7018)
* 1ef1d65bef2e2ecc2a2c04b2dcc99bc6a10f9e9c fix(cli): remove obsolete autopilot priority flag (#7557)
* 620f99d8ce168ee56bfcfacf359b3bcca30f21bf fix(codex): wait for authoritative turn completion (#7554)
* 3d47913c8526ccf669413a93d09e514a6d9b9657 fix(issues): allow verified autopilot child assignment (MUL-6161) (#7275)
* 7b8c7839913f8cd9bbb6574db609f3e7baea69d0 fix(issues): let verified autopilot runs assign private agents (#7572)
* 8442504201b302cfd6c40c7b8eb8a508bf254d0b fix(runtimes): hide private runtimes from non-owners (#7435)
* f74a71060fcdd45318cae6fb7caf7c0d7e71cdfa perf(web): preload landing hero background (MUL-6711) (#7582)
* 09a2410e882be8435bd6c4a26e03f7e288038203 refactor(agents): rename starter_prompts to conversation_starters (MUL-6709) (#7592)
* 55fe792483459585ab7d848066a34599851fde40 refactor(cloud): use URL-only internal connection (#7566)