agentsclimarketplace

Research watch

Skill kinhluan/skills/.agent-skills/research-watch/research-watch

πŸš€ Professional Multi-Agent Skills

Install
npx -y skills add kinhluan/skills --skill research-watch

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Monitor a research topic for new papers, trends, and breakthroughs. Set up recurring watches on arXiv, Semantic Scholar, or specific authors/venues. Use when tracking a fast-moving field, waiting for follow-up work, or maintaining a living literature review. Inspired by Feynman's /watch workflow.

SKILL.md

7.4 KB, as published. Nobody here has run it

Research Watch

Stay current without drowning in papers.

Set up intelligent monitoring for research topics, authors, venues, or keywords. Get notified when relevant new work appears, with automatic relevance scoring and synthesis.

"The literature moves fast. Your knowledge should move faster." β€” Anonymous Researcher


1. Watch Types

1.1 Topic Watch

Monitor a research topic for new papers:

Topic: "mechanistic interpretability"
Keywords: ["mechanistic interpretability", "circuit tracing", "activation patching"]
Venues: [NeurIPS, ICML, ICLR, arXiv cs.LG]
Frequency: weekly

Relevance scoring:

ScoreCriteriaAction
90-100Directly addresses your research questionRead immediately
70-89Related method or applicationRead within 2 days
50-69Same field, tangential topicSkim abstract
< 50Same broad areaSkip, but log for completeness

1.2 Author Watch

Track specific researchers:

Authors: ["Yann LeCun", "Yoshua Bengio", "Geoffrey Hinton"]
Rationale: "Foundational figures in deep learning"
Alert on: new papers, preprints, blog posts

1.3 Venue Watch

Monitor specific conferences/journals:

Venues: ["NeurIPS", "ICML", "ICLR", "JMLR"]
Filter: papers matching [your keywords]
Alert: when proceedings published

1.4 Citation Watch

Track who cites your work or key papers:

Target papers: [your-paper-id, foundational-paper-id]
Alert on: new citations
Include: citation context (how they cite you)

2. Setup Protocol

Step 1 β€” Define Watch Scope

Watch Name: [descriptive name]
Type: [topic / author / venue / citation]
Query: [keywords, authors, or paper IDs]
Sources: [arXiv / Semantic Scholar / Google Scholar / PubMed]
Frequency: [daily / weekly / monthly]

Step 2 β€” Configure Filters

Exclude noise:

  • Negative keywords: ["survey", "review", "tutorial"] (if you want only original research)
  • Minimum citations: [0 for new work, 10 for established work]
  • Date range: [last 30 days / last 90 days / all time]

Include priority:

  • CCF-A venues: always alert
  • arXiv from known authors: always alert
  • Preprints with >50 citations in first month: flag as potentially important

Step 3 β€” Set Up Alerts

arXiv RSS/API:

# arXiv API query
http://export.arxiv.org/api/query?search_query=all:mechanistic+interpretability&sortBy=submittedDate&sortOrder=descending&max_results=10

Semantic Scholar Alerts:

# Semantic Scholar search with date filter
curl "https://api.semanticscholar.org/graph/v1/paper/search?query=mechanistic+interpretability&fields=title,authors,year,abstract,citationCount&publicationDateOrYear=2024:2025"

Google Scholar Alerts:

Step 4 β€” Synthesis Template

When new papers arrive, produce a brief synthesis:

## Research Watch Report: [Topic] β€” [Date Range]

### New Papers (N found)

#### πŸ”΄ High Relevance
1. **[Title]** by [Authors] ([Venue], [Year])
   - **Why relevant:** [1 sentence]
   - **Key claim:** [1 sentence]
   - **Relation to your work:** [how it connects]
   - **Action:** [read / skim / save for later]

#### 🟑 Medium Relevance
2. ...

#### 🟒 Low Relevance
3. ...

### Trend Analysis
- **Emerging direction:** [what's new this period]
- **Confirming trend:** [what continues from before]
- **Contradicting result:** [paper that challenges consensus]

### Recommended Actions
1. [Read paper X immediately β€” directly relevant]
2. [Add paper Y to reading list β€” related method]
3. [Update SOTA matrix with paper Z β€” new baseline]

3. Watch Management

Active Watches Log

Keep a log of all active watches:

## Active Research Watches

| Watch Name | Type | Query | Frequency | Last Check | Papers Found |
|---|---|---|---|---|---|
| MechInterp-2024 | Topic | "mechanistic interpretability" | Weekly | 2024-01-15 | 12 |
| Transformer-Scaling | Topic | "scaling laws" + "transformer" | Weekly | 2024-01-15 | 8 |
| LeCun-Papers | Author | "Yann LeCun" | Daily | 2024-01-15 | 1 |
| NeurIPS-2024 | Venue | NeurIPS 2024 proceedings | Monthly | 2024-01-15 | 0 |

Watch Review Cycle

Monthly:

  • Review all active watches
  • Deactivate watches that are no longer relevant
  • Merge overlapping watches
  • Update keywords based on evolving research focus

Quarterly:

  • Assess watch effectiveness (signal-to-noise ratio)
  • Adjust relevance thresholds
  • Add new watches for emerging sub-topics

4. Integration with Research Workflow

Watch β†’ SOTA Survey

Research Watch (ongoing)
    ↓ (monthly synthesis)
SOTA Survey Update (sota-survey)
    ↓
Updated Literature Matrix
    ↓
Research Question Refinement (research-question)

Watch β†’ Experiment

Research Watch detects new baseline
    ↓
Paper Audit (paper-audit) β€” verify claims
    ↓
Add to experiment tracking (experiment-tracking)
    ↓
Run comparison experiment

Watch β†’ Paper

Research Watch detects related work published
    ↓
Update Related Work section (paper-writing)
    ↓
Add citation + differentiate your work

5. Tools for Research Watching

Manual (Free)

  • arXiv RSS: Subscribe to category feeds
  • Google Scholar Alerts: Email alerts for queries
  • Semantic Scholar: Weekly digest for followed authors
  • Twitter/X: Follow researchers (noisy but fast)

Semi-Automated

  • Zotero + RSS: Auto-import papers from feeds
  • Notion/Airtable: Track papers with custom views
  • Obsidian: Link papers in knowledge graph

Automated (Future)

  • LLM-powered filtering: Auto-score relevance, generate summaries
  • Slack/Discord bots: Push alerts to team channels
  • Auto-synthesis: Generate weekly reports automatically

6. Anti-Patterns

Anti-patternProblemFix
Too many watchesInformation overloadLimit to 3-5 active watches
Too broad queries90% irrelevant papersUse exact phrases, add negative keywords
Set-and-forgetWatches become staleMonthly review and cleanup
No synthesisPapers accumulate unreadWeekly synthesis, not just alerts
Ignoring low-relevanceMiss paradigm shiftsMonthly review of "skipped" papers

Integration with Other Skills

This skill providesRelated skillFor deeper dive
New paper discoverysota-surveyDeep analysis of discovered papers
Paper verificationpaper-auditCheck if new papers are trustworthy
Literature updatepaper-writingUpdate Related Work sections
Trend detectionresearch-questionIdentify new research directions

References

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.