You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Nov 14, 2026. It is now read-only.
Repository navigation
Commit e7461b4
Browse filesBrowse the repository at this point in the historyBrowse files
* feat(auth): add ALIBABA_CHINA_BASE_URL and base_url param to _apply_alibaba_config
* feat(auth): probe China endpoint on 401 before failing Alibaba login
* test(auth): add China-region fallback + workspace-key detection tests
* feat(auth): detect sk-ws- workspace keys and give targeted endpoint guidance
* feat(auth): interactive workspace endpoint prompt + update Alibaba model catalog
- Prompt for workspace endpoint host in TUI when sk-ws- key is entered,
eliminating the need to pre-set DASHSCOPE_BASE_URL for workspace keys
- Add base_url param to login_alibaba_api_key for caller-supplied endpoint
- Add qwen3.7-plus, qwen3-coder-plus, qwen3-coder-flash to catalog
- Remove kimi-k2.5, glm-5, MiniMax-M2.5 (absent from live endpoint)
- Correct qwen3.7-max context size to 1M (was 262k)
* chore(release): prepare 0.36.0
* fix(llm): use enable_thinking for Qwen models on DashScope
DashScope's OpenAI-compatible API uses extra_body={"enable_thinking": true/false}
for Qwen thinking models. Sending reasoning_effort triggers RouteError: Service
route not found (HTTP 500). Treat Qwen like Kimi K2/GLM: skip with_thinking()
and send the provider-specific toggle instead.
* fix(llm): use enable_thinking for all DashScope reasoning models
Extend the Qwen-only fix to cover all models accessed via DashScope
endpoints (kimi-k2.6, GLM, DeepSeek, etc.) using URL-based detection.
* fix(llm): use Moonshot thinking format for kimi-k2.6 on DashScope
kimi-k2.6 uses {"thinking": {"type": ...}} on all providers including
DashScope workspace; enable_thinking is Qwen-only. Remove the DashScope
exclusion from is_kimi_openai_legacy/is_glm_openai_legacy so the correct
Moonshot body format is applied regardless of endpoint.
* fix: harden Alibaba auth and agent guardrails
* docs(agents): support local agent instructions
Copy file name to clipboardExpand all lines: CHANGELOG.md
+8Lines changed: 8 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -15,6 +15,14 @@ GitHub Releases page; `0.8.0` is the new starting line.
15
15
16
16
## Unreleased
17
17
18
+
## 0.36.0 (2026-06-05)
19
+
20
+
-**Alibaba DashScope multi-region fallback.** Logging in with a China-region key (`dashscope.aliyuncs.com`) against the default US Virginia endpoint now auto-detects the mismatch and reconfigures for the correct endpoint rather than failing with a misleading "API key is wrong" error.
21
+
-**Alibaba Token Plan compatibility (`sk-ws-`).**`/login alibaba` now requires the dedicated workspace Base URL shown in the Token Plan console instead of accepting a public `/models` response as credential validation. Dedicated workspace endpoints hide Kimi K2.6 when Alibaba advertises it without a working route, and use non-streaming Chat Completions for DeepSeek V3.2 because those endpoints return an empty SSE stream. Kimi requests on other Alibaba routes use DashScope's `enable_thinking` parameter.
22
+
-**Alibaba model catalog refresh.** Added Qwen3.7 Plus (1M context), Qwen3 Coder Plus, and Qwen3 Coder Flash. Removed `kimi-k2.5`, `glm-5`, and `MiniMax-M2.5` (absent from the live endpoint). Corrected Qwen3.7 Max context window to 1M tokens.
23
+
24
+
Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.36.0`, or use the native installer for your platform from the [Releases page](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
25
+
18
26
## 0.35.0 (2026-06-04)
19
27
20
28
-**Alibaba DashScope provider and MiniMax M3 catalog.**`/login alibaba` now configures Alibaba Cloud Model Studio / DashScope Token Plan models, including workspace-compatible endpoints and native GLM thinking behavior. MiniMax API-key login now defaults to MiniMax M3 with its larger context and multimodal capabilities.
Copy file name to clipboardExpand all lines: README.md
+22-22Lines changed: 22 additions & 22 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -50,13 +50,13 @@ It speaks the [**Agent Client Protocol (ACP)**](https://github.com/agentclientpr
50
50
51
51
---
52
52
53
-
## 🆕 What's New in 0.35.0
53
+
## 🆕 What's New in 0.36.0
54
54
55
-
-**`/stats` usage dashboard.**New slash command opens an interactive TUI showing token and cost breakdown by provider/model across Today / This Week / Last Week / All Time. Powered by a static pricing table and a session collector that walks `~/.pythinker/sessions/` wire files.
56
-
-**Z AI provider auth.**Login/logout via API key, model discovery, and OAuth selector wired into the TUI and `refresh_managed_models`.
57
-
-**Moonshot provider auth.**Login/logout via API key, model discovery (Kimi K2.x catalog), OAuth selector wired into the TUI and `refresh_managed_models`.
55
+
-**Alibaba DashScope multi-region fallback.**China-region keys now auto-detect the endpoint mismatch and reconfigure correctly instead of showing a misleading "API key is wrong" error.
56
+
-**Alibaba Token Plan compatibility (`sk-ws-`).**`/login alibaba` now asks for the dedicated workspace endpoint, avoids unroutable Kimi entries on those endpoints, and uses DeepSeek V3.2's working non-streaming mode.
57
+
-**Alibaba model catalog refresh.**Qwen3.7 Plus, Qwen3 Coder Plus, and Qwen3 Coder Flash added; deprecated `kimi-k2.5`, `glm-5`, and `MiniMax-M2.5` removed.
58
58
59
-
Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.35.0`, or use the native installer for your platform from the [Releases page](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
59
+
Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.36.0`, or use the native installer for your platform from the [Releases page](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
60
60
61
61
62
62
---
@@ -146,7 +146,7 @@ matches your OS — no Python, Node, or `uv` prerequisite.
Copy file name to clipboardExpand all lines: docs/en/guides/getting-started.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -44,7 +44,7 @@ On Windows, run the PowerShell bootstrap. It downloads the native installer, ver
44
44
irm https://pythinker.com/install.ps1 | iex
45
45
```
46
46
47
-
You can also download `PythinkerSetup-0.35.0.exe` manually from the [latest release](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
47
+
You can also download `PythinkerSetup-0.36.0.exe` manually from the [latest release](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
Copy file name to clipboardExpand all lines: docs/en/release-notes/changelog.md
+8Lines changed: 8 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -17,6 +17,14 @@ GitHub Releases page; `0.8.0` is the new starting line.
17
17
18
18
## Unreleased
19
19
20
+
## 0.36.0 (2026-06-05)
21
+
22
+
-**Alibaba DashScope multi-region fallback.** Logging in with a China-region key (`dashscope.aliyuncs.com`) against the default US Virginia endpoint now auto-detects the mismatch and reconfigures for the correct endpoint rather than failing with a misleading "API key is wrong" error.
23
+
-**Alibaba Token Plan compatibility (`sk-ws-`).**`/login alibaba` now requires the dedicated workspace Base URL shown in the Token Plan console instead of accepting a public `/models` response as credential validation. Dedicated workspace endpoints hide Kimi K2.6 when Alibaba advertises it without a working route, and use non-streaming Chat Completions for DeepSeek V3.2 because those endpoints return an empty SSE stream. Kimi requests on other Alibaba routes use DashScope's `enable_thinking` parameter.
24
+
-**Alibaba model catalog refresh.** Added Qwen3.7 Plus (1M context), Qwen3 Coder Plus, and Qwen3 Coder Flash. Removed `kimi-k2.5`, `glm-5`, and `MiniMax-M2.5` (absent from the live endpoint). Corrected Qwen3.7 Max context window to 1M tokens.
25
+
26
+
Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.36.0`, or use the native installer for your platform from the [Releases page](https://github.com/Pythoughts-labs/pythinker-code/releases/latest).
27
+
20
28
## 0.35.0 (2026-06-04)
21
29
22
30
-**Alibaba DashScope provider and MiniMax M3 catalog.**`/login alibaba` now configures Alibaba Cloud Model Studio / DashScope Token Plan models, including workspace-compatible endpoints and native GLM thinking behavior. MiniMax API-key login now defaults to MiniMax M3 with its larger context and multimodal capabilities.
Copy file name to clipboardExpand all lines: src/pythinker_code/agents/default/code_reviewer.yaml
+2Lines changed: 2 additions & 0 deletions
Original file line number
Diff line number
Diff line change
@@ -26,6 +26,8 @@ agent:
26
26
- Build a review context packet: base ref/diff scope or Reviewflow feature IDs, changed behavior, likely tests, user-visible impact, valid evidence paths, omitted/truncated context, and validation evidence.
27
27
- Flag only issues introduced or made reachable by the diff.
28
28
- Prefer no finding over vague speculation. Every finding must cite concrete evidence and a failure mode.
29
+
- Run the production guardrail gate before finalizing: check for cache stampedes, connection/resource leaks, missing boundary schemas, unhandled race conditions, naive retry loops, unbounded event callbacks/listeners, and IDOR/tenant-scope mistakes.
30
+
- Treat missing `finally` cleanup, absent schema validation at trust boundaries, unprotected shared-state mutation, non-jittered immediate retries, or identity from mutable client parameters as reject-level findings when reachable in the changed code.
29
31
30
32
Freshness check (run BEFORE flagging third-party library or framework misuse):
31
33
- For every third-party API, SDK call, framework primitive, or "best practice" the diff turns on, verify the current canonical usage. Prefer a context7 MCP query (e.g. `mcp__context7__query-docs` with the library id) when registered with the parent runtime; otherwise use `SearchWeb` to locate the official documentation and `FetchURL` to read the current page.
0 commit comments