Relevant jobs
AI / Algorithm Engineer
Full-stack / Indie Developer
Backend Engineer
Data Analyst / Data Engineer
nexus-agentjava-up-up/nexus-agent
A production-ready Agentic RAG agent implementation for the Java stack that upgrades a chatbot into a system that reason…
No ratings yet
Rate this skill
Stars657
Forks101
Watchers657
Open Issues3
InstallOne-line install
UnconfirmedWhether this repository follows the Agent Skills spec (that is, whether it contains a SKILL.md) has not been verified yet. Until that is settled, clone it and use what you need:
git clone https://github.com/java-up-up/nexus-agent.git
OverviewOverview
Key Features
A production-ready Agentic RAG agent implementation for the Java stack that upgrades a chatbot into a system that reasons, retrieves evidence, and calls tools. Key capabilities include: 1. ReAct Agent full implementation: supports web search, tool calling, multi-step reasoning, and Checkpoint persistence, so long-running tasks keep state and can resume after interruption; 2. Three-tier executor system: performs deterministic orchestration first and only routes complex questions into the agent reasoning loop, cutting unnecessary model calls; 3. Neo4j document structure graph: stores chapter hierarchy and citation relations in a graph database so retrieval can locate specific passages instead of whole documents; 4. Three-stage knowledge routing funnel: narrows candidate knowledge domains step by step, coarse filtering before fine ranking, reducing cross-domain false recalls; 5. Shadow routing quality observation: records routing decisions in parallel without affecting the live main path, for comparison and tuning; 6. RAG pre-orchestration engine: completes query rewriting, intent recognition, and retrieval strategy selection before retrieval; 7. Dual-channel hybrid retrieval: runs vector and keyword retrieval in parallel and fuses the ranked results, balancing semantic similarity and exact term matching; 8. Evidence budget control and no-evidence short-circuit: caps the number and length of evidence per answer and short-circuits when evidence is insufficient, preventing fabricated answers; 9. Parent-Child chunk aggregation: matches on small chunks and backfills context with large chunks, balancing retrieval precision and generation completeness; 10. Three session memory strategies: choose different memory retention modes per scenario to control context length and information continuity in multi-turn dialogue.
TopicsTopics
RatingUser ratings
Nobody has rated this skill yet. Yours would be the first on the site — your vote becomes the initial average.
RepositoryRepository
| Skill ID | 1283 |
|---|---|
| Data source | GitHub |
| Repository | java-up-up/nexus-agent |
| Category | AI & Agent Skills / Coding & Engineering Agent |
| Primary language | Java |
| License | Apache License 2.0 |
| Default branch | master |
| GitHub repository ID | 1214166855 |
| Node ID | R_kgDOSF63Rw |
| Rank | #1038 (position within the source ranking) |
| Site-wide popularity rank | #1192 (computed live from Stars, across all 5833 entries) |
GrowthStar / Fork growth
One snapshot per day, tracking how Stars and Forks changed
- Now657
- Previous point · 09-29 648
- vs previous point +9
- 2 data points 12 days
- Now101
- Previous point · 09-29 100
- vs previous point +1
- 2 data points 12 days
TimelineTimeline
Repository created2026-04-18 07:50:437 months ago
Data last updated2026-10-26 12:07:5723 days ago
Last commit2026-09-16 08:25:412 months ago
CommitLatest commit
Version [latest] committed by shining-stars-l. Update Category: Refactoring & Performance Optimization. Commit details: 优化介绍内容. This update optimizes execution logic and refactors codebase structure.