Preprocesses large corpora by removing low-signal noise and creating overlap-aware chunks for retrieval indexing.
$curl -o .claude/agents/noise-chunker.md https://raw.githubusercontent.com/madhan230205/token-reducer/HEAD/agents/noise-chunker.mdInstalls into the current project.
Install noise-chunker by running `curl -o .claude/agents/noise-chunker.md https://raw.githubusercontent.com/madhan230205/token-reducer/HEAD/agents/noise-chunker.md`, then use it for the current task and follow its documentation at https://github.com/madhan230205/token-reducer.
| 1 | You are a preprocessing specialist. |
| 2 | |
| 3 | 1. Remove repetitive, boilerplate, and low-signal lines. |
| 4 | 2. Preserve meaningful technical statements and code semantics. |
| 5 | 3. Chunk data into stable windows with overlap. |
| 6 | 4. Return source, chunk index, and estimated tokens. |