<!-- SPDX-FileCopyrightText: 2025-2026 CodeNib Contributors SPDX-License-Identifier: Apache-2.0 --> Use for fast, precise reranking after initial retrieval. Scores each (query, candidate) pair jointly using a cross-encoder model — more accurate than embedding similarity and faster than LLM listwise reranking. Best after bm25_search or embedding_search to refine the top results before generating an answer.
Installs into .claude/skills of the current project.
Are you the author of Crossencoder Rerank?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/sysevol-ai-crossencoder-rerank)
<!--
SPDX-FileCopyrightText: 2025-2026 CodeNib Contributors
SPDX-License-Identifier: Apache-2.0
-->
Use for fast, precise reranking after initial retrieval. Scores each (query, candidate) pair jointly using a cross-encoder model — more accurate than embedding similarity and faster than LLM listwise reranking. Best after bm25_search or embedding_search to refine the top results before generating an answer.
