Design retrieval-augmented generation pipelines with hybrid retrieval, reranking, citations, and refusal behavior.
npx skills add https://github.com/f4rkh4d/forge-skill --skill forge-rag
Please help me install this Agent Skill. Skill: forge-rag Source: https://github.com/f4rkh4d/forge-skill/tree/main/skills/llm/forge-rag Command: npx skills add https://github.com/f4rkh4d/forge-skill --skill forge-rag
It prevents RAG systems from hallucinating and returning unverifiable answers by enforcing retrieval quality, chunking structure, citation discipline, and refusal behavior.
Ask your AI to redesign your RAG pipeline to use semantic hierarchical chunking, hybrid BM25+dense retrieval with reranking, and a generation prompt that requires inline citations to verified retrieved chunks or returns a refusal when no good passages are found.
High-intent search queries and answers about installing and using this skill.
To stop retrieval-augmented generation hallucinations, enforce low-confidence refusal paths and require inline citations to verified retrieved chunks so the model avoids fabrication when passage relevance is insufficient.
The best way to combine BM25 and dense retrieval for RAG is using reciprocal rank fusion to merge results, followed by cross-encoder reranking to select the most relevant passages for generation.
Semantic hierarchical chunking improves retrieval accuracy by chunking documents into meaningful units with overlap to preserve document structure, ensuring the model answers with properly contextualized information.
Evaluate a RAG system using an eval-driven regression testing set that includes both answerable queries and refusal cases, ensuring the pipeline maintains citation discipline and retrieval quality across updates.
A RAG system returns answers without relevant passages due to missing low-confidence refusal guardrails; implementing verified citations and generation limits forces the model to refuse rather than fabricate.
Scheduled indexing works with version-aware retrieval to maintain knowledge base freshness, automatically updating document chunks while preserving retrieval quality and citation discipline across versions.
yu13130122297Identify uncertainties in research framing, datasets, metrics, and baselines.
rivaldiekaptrrrDefine and run eval suites to validate AI task performance.
latestaiagentsCombine dense vector embeddings with sparse BM25 retrieval for RAG systems.
JCETools-PetraDesign RAG pipelines with chunking, embedding, retrieval, and evaluation metrics.
