@@ -1,79 +1,6 @@
11 # TODO
22
3-- Most models hang for long periods during turns. Sometines needing the user to interrupt to continue. This happened in recent sessions in the db. This is not a provider issues, both opencode and pi using the same provider and model don't have hangups.
4-
5-- Tool hooks output prints inline with spinner when they should output in a new line and indented like the tool calls output.
6-- add a spinner state showing the user when the undo snapshotting is happening
7-- Structuted output when a skill is auto loaded for the agent
8-- Kimi-k2.5, GLM-5 and other models that use the same sdk produce these errors:
9-
10-Error 1:
11-
12-```
13-"{\"error\":{\"object\":\"error\",\"type\":\"invalid_request_error\",\"code\":\"invalid_request_error\",\"message\":\"Invalid tool call in messages: tool_calls[].function.arguments for function 'read' must decode to a JSON object, got str\"}}",
14-```
15-
16-Error 2:
17-
18-```
19-2543 | text: responseBody,
20-2544 | schema: errorSchema
21-2545 | });
22-2546 | return {
23-2547 | responseHeaders,
24-2548 | value: new APICallError4({
25- ^
26-AI_APICallError: Expected dict in tool call arguments, got <class 'str'>
27- cause: undefined,
28- url: "https://opencode.ai/zen/v1/chat/completions",
29- requestBodyValues: {
30- model: "glm-5",
31- user: undefined,
32- max_tokens: undefined,
33- temperature: undefined,
34- top_p: undefined,
35- frequency_penalty: undefined,
36- presence_penalty: undefined,
37- response_format: undefined,
38- stop: undefined,
39- seed: undefined,
40- reasoning_effort: undefined,
41- verbosity: undefined,
42- messages: [
43- [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Objec …
44- ],
45- tools: [
46- [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...], [Object ...]
47- ],
48- tool_choice: "auto",
49- stream: true,
50- stream_options: undefined,
51-},
52- statusCode: 400,
53- responseHeaders: {
54- "cf-ray": "9dc2e9f38ed09063-BOS",
55- connection: "keep-alive",
56- "content-type": "application/json",
57- date: "Sat, 14 Mar 2026 11:24:46 GMT",
58- server: "cloudflare",
59- "transfer-encoding": "chunked",
60-},
61- responseBody: "{\n\t\"error\": {\n\t\t\"code\": 400,\n\t\t\"message\": \"Expected dict in tool call arguments, got \\u003cclass 'str'\\u003e\",\n\t\t\"type\": \"Bad Request\"\n\t}\n}\n",
62- isRetryable: false,
63- data: {
64- error: [Object ...],
65-},
66- vercel.ai.error: true,
67- vercel.ai.error.AI_APICallError: true,
68-
69- at /home/xonecas/src/mini-coder/node_modules/@ai-sdk/provider-utils/dist/index.mjs:2548:18
70- at /home/xonecas/src/mini-coder/node_modules/@ai-sdk/provider-utils/dist/index.mjs:2391:34
71- at processTicksAndRejections (native:7:39)
72-
73-✖ API error 400
74- https://opencode.ai/zen/v1/chat/completions
75-```
76-
3+- read/write tool call output only visible when tool call response is already visible too.
774 - Investigate db lock issues with spawning parallel subagents:
785
796 ```
@@ -86,15 +13,6 @@ AI_APICallError: Expected dict in tool call arguments, got <class 'str'>
8613 ◆ Subagent parallelism hit a SQLite lock in the shared app DB, so I’m finishing the audit with direct reads and a smaller follow-up pass grounded in file references.
8714 ```
8815
89-- double line print bug:
90-
91-```
92-· reasoning
93-│ Mapping command definitions
94-◆ Progress update: I’ve loaded the idea/TODO and core user docs; next I’m scanning implementation files to map actual behavior against t
95-◆ Progress update: I’ve loaded the idea/TODO and core user docs; next I’m scanning implementation files to map actual behavior against those sources and capture precise gaps/inconsistencies for the report.
96-```
97-
9816 - read tool path issues, says file not found on valid path:
9917
10018 ```
@@ -108,6 +26,9 @@ AI_APICallError: Expected dict in tool call arguments, got <class 'str'>
10826 · src/cli/stream-render.ts 296 lines
10927 ```
11028
29+- add a spinner state showing the user when the undo snapshotting is happening
30+- Structuted output when a skill is auto loaded for the agent and improve context pruning output to match out style
31+
11132 ---
11233
11334 ## UI Audit