Task set v1.0 · China local
ARTB-C03 · v1.0
Xiaohongshu research
The situation
Before a trip, I want recent real notes from Xiaohongshu.
What we say to the agent
Find 5 Xiaohongshu notes from the past month about {topic}, with links, and summarize the common tips.
Changes every round
{topic}
Prompt templates are public. The concrete values are published with each round's results.
How it is judged
Pass
- 5 real notes, on topic, within 30 days, links open.
Partial
- 3-4 meet every requirement.
Fail
- 2 or fewer, or any invented note.
Evidence we keep
screen_recording, link_check_log
Results
| Doubao | Not tested |
| Kimi | Not tested |
| Yuanbao | Not tested |
| Coze Space | Not tested |
| AutoGLM | Not tested |
| ChatGPT agent | Not tested |
| Gemini | Not tested |
| Manus | Not tested |
| OpenClaw | Not tested |
| Hermes Agent | Not tested |