fix: stop prepending reasoning_content to model output

[li-xiu-qi]

Problem

Fixes #62

When using cloud reasoning models (e.g. step-3.7-flash) via Chat Completions API, the parser fails on the first step:

ValueError: ui_action must contain 'action' or 'action_type'. Got keys: ['cot']

Root Cause

ask_llm_v2.py detects reasoning_content in the API response and wraps it as a <think> block, prepending it to the model's content:

result = "<think>" + reasoning + "</think>" + "\n" + llm_content

But the model's content already contains <THINK> tags (required by the parser prompt), so this creates two layers:

<think>reasoning_content from API</think>       ← prepended by ask_llm_v2.py
<THINK>model's own thinking</THINK>             ← from model content
explain:...  action:CLICK  point:500,300  ...   ← parser needs this

The parser's str2action() uses split("</THINK>")[1] to extract key-value pairs, which grabs content after the first </THINK> — i.e., the second THINK block text, not the action parameters.

Three API Comparison

API	Reasoning Location	Two-layer THINK?
Chat Completions	reasoning_content field → prepended by ask_llm_v2.py	Yes
Messages (Anthropic)	thinking type block (separate)	No
Responses	reasoning type output item (separate)	No

The model's content always has exactly one <THINK> layer. The problem is purely in ask_llm_v2.py's handling.

Fix

Stop prepending reasoning_content to the result. Log it for debugging, but return content as-is. The content already follows the expected <THINK>...</THINK>\texplain:...\taction:...\tsummary:... format.

Only tools/ask_llm_v2.py is modified (1 file, 19 insertions, 26 deletions).

Verification

Tested with step-3.7-flash controlling a real Android device (iQOO Z9 Turbo+):

Task	Steps	Result
Open Settings > About Phone > read model info	5 steps	Completed successfully
Open WeChat > find and open specific group chat	5 steps	Completed successfully

Both tasks failed before the fix (parser error on step 1) and succeeded after.