ResearchMonitor

**Judgement:** ResearchMonitor is a solid, focused skill for ongoing literature surveillance in specific research areas. It’s well-aligned with workflows where the assistant runs periodically, checks multiple scholarly sources, tracks what has already been seen, and only surfaces genuinely new items. **Strengths & Benefits:** - Automates repetitive web and literature searches for defined topics. - Uses `research_config.json` plus helper scripts to remember topics, last-checked time, and already-seen items. - Reduces user noise by explicitly avoiding notifications when nothing significant is new. - Produces concise, markdown-formatted briefings with title, source, and a one-sentence relevance summary. **Risks & Limitations:** - Dependent on generic `search_web` quality and external sites (arXiv, IEEE Xplore, Google Scholar, X/Twitter); changes or blocks can degrade coverage or cause missed items. - The definition of “significantly new” is heuristic (recent publication dates and seen-checks), so important items can occasionally be missed or delayed. - Performance and latency may degrade when monitoring many topics or when search results are very large. - Accuracy of relevance judgments depends on how well the model interprets search snippets and metadata. **Recommended Scenarios:** - Academics, students, and industrial researchers who want low-friction, incremental updates in a few well-defined areas. - Users who prefer not to be interrupted unless there is genuinely new literature or event information. - Research workflows where daily or near-daily background monitoring is useful, and occasional gaps or misses are acceptable. **Less Suitable For:** - Mission-critical systematic reviews that demand exhaustive, auditable coverage of all publications. - Users needing fine-grained control over ranking, source selection, or complex filtering beyond simple topic keywords.