noise-reduction

Diagnose and reduce noise in agent-user conversation transcripts.

64|5|Updated Mar 13, 2026
One-click install
npx skills add https://github.com/MattWenJun/ClaWiser --skill noise-reduction
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: noise-reduction
Source: https://github.com/MattWenJun/ClaWiser/tree/main/assets/noise-reduction
Command: npx skills add https://github.com/MattWenJun/ClaWiser --skill noise-reduction

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Noise in conversation transcripts degrades memory recall and makes it hard to retrieve relevant context. This Skill provides structured noise profiling, metadata stripping, and conservative filtering to preserve meaningful user-agent content while removing noise patterns.

Core Features & Use Cases

  • Noise profiling and pattern inventory for agent-user transcripts across channels.
  • Metadata stripping and rule-based filtering to reduce irrelevant content without losing essential context.
  • Validation and reporting to quantify compression, false positives, and missed noise for continuous improvement.

Quick Start

Run diagnose-noise.js <YYYY-MM-DD> to generate an initial noise profile for that date.

Frequently Asked Questions about noise-reduction

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce noise in conversation transcripts to improve memory recall?

To reduce conversation noise, this Skill profiles transcript patterns and applies rule-based filtering to strip metadata wrappers, preserving meaningful content for better memory recall.

Can I clean Telegram, Discord, and Slack transcripts with noise profiling?

Yes, noise profiling works across OpenClaw sessions and transcripts from channels like Telegram, Discord, and Slack, filtering irrelevant messages to improve transcript quality.

How do I diagnose noise patterns in agent-user conversations?

You diagnose noise patterns by running the validation script diagnose-noise.js with a target date, generating a structured noise profile that catalogs irrelevant content.

What is the best way to strip metadata wrappers from chat transcripts without losing context?

The best way to strip metadata wrappers safely is using conservative rule-based filtering, which removes irrelevant noise patterns while implementing safeguards to preserve essential context.

Does noise reduction filtering cause false positives in data cleaning?

Noise reduction includes validation and reporting scripts that explicitly quantify compression, false positives, and missed noise, ensuring false positives are tracked for continuous improvement.

When should I not use rule-based filtering for transcript data cleaning?

You should avoid rule-based transcript filtering when conversations contain unstructured noise patterns that fall outside the diagnosed profile, potentially triggering conservative safeguards and missing noise.