| Age | Commit message (Expand) | Author |
|---|
| 2026-06-03 | fix: warm the SAME Anthropic message-cache bucket as real turns | Adam Malczewski |
| 2026-06-03 | feat: prompt cache warming for idle tabs | Adam Malczewski |
| 2026-06-02 | feat(chat): paste-to-attach images/PDFs with model capability check | Adam Malczewski |
| 2026-06-02 | feat(agents): per-model reasoning effort level | Adam Malczewski |
| 2026-06-01 | fix(queue): consume queued messages after a turn ends (start a new turn) | Adam Malczewski |
| 2026-05-30 | fix(agent): stream thinking for all adaptive Claude models, not just Opus 4.7 | Adam Malczewski |
| 2026-05-30 | chore(notes): collect loose root docs into notes/; add reconcile edge-cases note | Adam Malczewski |
| 2026-05-30 | refactor(chunks): append-only chunk log with per-step cache-stable wire | Adam Malczewski |
| 2026-05-30 | feat(cache): Anthropic prompt caching, usage telemetry, and Cache Rate view | Adam Malczewski |
| 2026-05-29 | fix: handle unavailable tool calls via native v6 tool-error event, not synthe... | Adam Malczewski |
| 2026-05-28 | refactor(core): upgrade ai-sdk v4 → v6 + Anthropic/openai-compatible reason... | Adam Malczewski |
| 2026-05-27 | refactor: ChatMessage.chunks[] union — interleaved thinking, tool batching,... | Adam Malczewski |
| 2026-05-22 | feat: agent summoning system, todo improvements, security fixes, double-execu... | Adam Malczewski |
| 2026-05-19 | feat: Phase 2 — shell permissions, tree-sitter analysis, permission UI | Adam Malczewski |
| 2026-05-19 | fix: DeepSeek reasoning_content dropped on multi-step tool calls | Adam Malczewski |
| 2026-05-19 | Phase 1: single agent + basic UI | Adam Malczewski |
