From ai-frontier
AI frontier plugin — gives Claude and agents access to latest scientific research, SOTA discovery, and architecture evaluation. Use when needing SOTA methods, research consensus, implementation guidance, or evaluating whether a codebase uses optimal patterns and architectures.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ai-frontier:using-ai-frontierThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You have access to deep research intelligence via four data sources and four synthesis agents.
You have access to deep research intelligence via four data sources and four synthesis agents.
Citation count measures influence, not currency. A paper with 5000 citations from 2022 may be foundational but no longer SOTA. Always consider:
When unsure whether a method is still SOTA, use perplexity-search --sota to get web-grounded current status.
When you encounter these situations, use the appropriate tool:
Agent: architecture-evaluator
Triggers: "Is this the right architecture?", "Evaluate my approach", "Are there better patterns for X?", "Review this against SOTA", "Is my architecture optimal?", "What patterns should I use for X?"
Returns: SCOPE, CURRENT ARCHITECTURE (with ASCII diagram), SOTA ALTERNATIVES, GAP ANALYSIS, RECOMMENDATIONS, REFERENCES
Agent: literature-reviewer
Triggers: "What's the SOTA for X?", "Survey of approaches to X", "What does the research say about X?"
Returns: DATA SOURCES, CONSENSUS, FRONTIER, OPEN QUESTIONS, KEY PAPERS, METHOD TAXONOMY, APPLICABILITY
Agent: method-analyst
Triggers: "Should I use X or Y?", "Compare methods for Z", "Tradeoffs between approaches"
Returns: COMPARISON MATRIX, RECOMMENDATION, IMPLEMENTATION NOTES
Agent: implementation-guide
Triggers: "How to implement X from paper Y?", "Code for method Z", "Architecture of approach X"
Returns: CORE ALGORITHM, ARCHITECTURE, REFERENCE IMPLEMENTATIONS, ADAPTATION GUIDE
| Need | Skill | Example |
|---|---|---|
| Bleeding-edge preprints | arxiv-search | Latest papers on speculative decoding |
| Citations + TLDRs + impact | semantic-scholar-search | How influential is paper X? |
| AI summaries + trending | hf-papers-search | Trending papers, AI-extracted keywords |
| Current SOTA / what replaced X | perplexity-search --sota | What replaced ReAct for agent reasoning? |
| Last N days developments | perplexity-search --recent | New RAG frameworks this month |
| Web-grounded fact check | perplexity-search | Is paper X still considered SOTA? |
All retrieval skills return a unified JSON envelope. Agents return structured markdown sections.
--sota) → what's current, what's supersededIf Perplexity is unavailable (no API key), agents still work with academic data only — results will have citation-based ranking without web-grounded SOTA validation.
npx claudepluginhub flight505/ai-frontierResearches SOTA AI/ML literature for topics, methods, or architectures; finds papers, builds comparison tables, recommends codebase strategies, and generates phased implementation plans.
Multi-agent research skill for technology evaluation, SOTA analysis, codebase archaeology, and competitive analysis. Uses wave-based pattern to gather breadth then synthesize.
Guides research planning: analyzes query complexity, decomposes via Self-Ask, Least-to-Most, DAG, parallel strategies; scales effort, sets stopping criteria, avoids anti-patterns.