A PCOnline journalist had ArkClaw set up three research agents and one document lead that together produced a research report on OpenClaw, and also scheduled a daily market data summary. The reporter noted the cloud version cannot read and write local files the way a local install can.
- First reported
- pconline.com.cn · 13 Mar 2026
- Added
- 11 Oct 2026
- Source checked
- 11 Oct 2026
What they did
A PCOnline journalist connected ArkClaw to Feishu and added a market-data search skill, then asked it to set up a project that gathers the day's limit-up and limit-down stocks, sector moves and other A-share data after the close and writes a brief to a Feishu doc at 16:45 every trading day. In a separate test, the reporter installed two team-building skills and had ArkClaw spin up four agents: three researchers covering basics, technology and competitors, plus a document lead that merged their findings into one report and sent it to a colleague as a Feishu doc.
What happened
The journalist says ArkClaw built the scripts and schedule itself, and the first brief arrived the same day, with later ones archived in Feishu. The four-agent run produced a structured research report on OpenClaw.
Keep in mind
The reporter also asked for next-day stock predictions, which is investment territory and not something to rely on. The article reads partly as a promotion, and it notes the cloud version cannot read or write files on your own computer the way a local install can.
Try it with your agent
Every working day at <time>, collect <the market or industry data you track> from <your data sources>, and write a one-page brief with the main moves, notable outliers and the sources used. Save it to <your shared doc folder> with the date in the title. Only report what the data shows; do not make forecasts.
Replace the <angle brackets> with your own details.
Original source
pconline.com.cnHands-on reviewA journalist or reviewer used the agent and wrote up what happened.
Tools for this kind of work
- RepoExtract tables and text from complex or scanned PDFs with Docling
- RepoSearch the web and scrape sites into clean data with Firecrawl
