Merge remote-tracking branch 'upstream/main'

# Conflicts:
#	README.md
#	uv.lock
This commit is contained in:
m4
2026-07-13 09:46:07 +08:00
78 changed files with 3718 additions and 3766 deletions
+2 -4
View File
@@ -179,6 +179,7 @@ only a transport event, not proof that the LangGraph run has finished.
<details>
<summary>📦 Release Highlights — version changelog</summary>
- **[11 Jul 2026]** **[v0.2.2](https://github.com/EvoScientist/EvoScientist/releases/tag/v0.2.2)** — New models selectable in onboarding and `/model`: GPT-5.6 (sol, terra, luna) for OpenAI and OpenRouter, plus Grok 4.5 and Tencent Hunyuan HY3 on OpenRouter; tighter config-file permissions and a reworked onboarding OAuth flow for auxiliary models.
- **[07 Jul 2026]** **[v0.2.2](https://github.com/EvoScientist/EvoScientist/releases/tag/v0.2.2)** — WebUI streaming resilience hotfix: SSE disconnects no longer imply run completion; the composer stays on **Stop** until backend thread state is terminal; final-answer checkpoint recovery fills dropped response tails; tool-selection JSON payloads are filtered from live transcripts; `EVOSCIENTIST_WEBUI_PACKAGE` lets the launcher run a local patched WebUI package for validation before npm publication.
- **[05 Jul 2026]** **[v0.2.1](https://github.com/EvoScientist/EvoScientist/releases/tag/v0.2.1)** — AutoSkills: EvoMemory drafts reusable skills from its own observation clusters for you to review via `/autoskills`; a new `--output-format stream-json` for headless / SDK clients; richer slash-command completions; Windows UTF-8 config reads; a TUI welcome-banner fix; langchain-openrouter 0.2.5.
- **[26 Jun 2026]** **[v0.2.0](https://github.com/EvoScientist/EvoScientist/releases/tag/v0.2.0)** — Scheduled tasks: cron-style recurring runs via `/schedule` or natural language, run unattended with shell-access gating; self-linking memory that connects observations into a knowledge graph (complements / contradicts / supersedes); a read-only `GET /api/models` endpoint for the WebUI model picker.
@@ -463,10 +464,7 @@ EvoSci deploy # standalone LangGraph server for external UIs
EvoSci -p "query" --output-format stream-json --auto-mode # JSONL event stream on stdout (for programmatic clients)
```
`--output-format stream-json` makes a single-shot (`-p`) run emit its native
events as line-delimited JSON on stdout (one object per line), with all human
output on stderr — the integration surface for headless clients (e.g. an agent
runtime). See [docs/stream-json.md](docs/stream-json.md) for the event schema.
`--output-format stream-json` makes a single-shot (`-p`) run emit its native events as line-delimited JSON on stdout (one object per line), with all human output on stderr — the integration surface for headless clients (e.g. an agent runtime). See [docs/guides/stream-json.md](docs/guides/stream-json.md) for the event schema.
</details>