skills.rest
Lets7512
Lets7512COMMUNITY
Explore all skills by this author →
GitHub

rlm

Analyze large files via a six-step Python protocol that prints only summaries.

24|3|Updated Mar 5, 2026
Search Tags:#data-analysis#token-efficiency#log-analysis#codebase-analysis#large-context#rlm
One-click install
npx skills add https://github.com/Lets7512/rlm-skill --skill rlm-lets7512
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rlm
Source: https://github.com/Lets7512/rlm-skill/tree/main/skills/rlm
Command: npx skills add https://github.com/Lets7512/rlm-skill --skill rlm-lets7512

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Massive data and files often exceed a single context window, forcing manual, error-prone extraction. The RLM protocol enables structured external processing and prints only summaries into the context.

Core Features & Use Cases

  • 6-step protocol (METADATA, PEEK, SEARCH, ANALYZE, SYNTHESIZE, SUBMIT) executed via small Python snippets to avoid loading raw data into context.
  • Sub-agent decomposition for large chunks to parallelize analysis without leaking data into the prompt or context.
  • Safe, deterministic execution: all heavy lifting happens outside the model context; only concise results enter context.
  • Use cases include processing massive logs, scanning large codebases, and analyzing datasets that exceed token limits.

Quick Start

Provide a large file path and run the RLM workflow to start processing with METADATA, PEEK, SEARCH, ANALYZE, SYNTHESIZE, and SUBMIT.

Frequently Asked Questions about rlm

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

FAQPage Schema
How do I analyze large log files that exceed the LLM context window?▼

To analyze large log files exceeding the context window, you can use a multi-step protocol that extracts metadata, peeks, searches, and synthesizes findings. This ensures raw data never enters the model context, printing only concise summaries.

What is the best way to audit a large codebase without flooding the context?▼

Auditing a large codebase without context flooding is best achieved through a structured external processing protocol. It executes defined steps via Python snippets and sub-agents, ensuring only selective outputs and summaries enter the context.

Can I process massive datasets in parallel without leaking raw data into the prompt?▼

Yes, you can process massive datasets without leaking raw data by using sub-agent decomposition. This approach parallelizes analysis of large chunks outside the model context, ensuring safe, deterministic execution and preventing context overload.

How does the RLM protocol work for token-efficient data analysis?▼

The RLM protocol works for token-efficient data analysis by orchestrating a six-step workflow: METADATA, PEEK, SEARCH, ANALYZE, SYNTHESIZE, and SUBMIT. It uses small Python snippets to process data externally, printing only actionable results into the context.

Related Skills

Explore More Software Engineering Skills→
nicholasgriffintnnicholasgriffintn
2

security-review

Identify security risks in codebases during reviews.

Community
Basic
softsparksoftspark
161

orchestrate

Coordinate multiple specialized agents to execute complex tasks in parallel.

Community
Advanced
HaykTarkhanyanHaykTarkhanyan

scholar-evaluation

Evaluate scholarly work with dimension-based scoring and weighted quality assessments.

Community
Advanced
parcadeiparcadei
3.9k

no-task-output

Enforce synchronous Task tool usage to prevent TaskOutput context flooding.

Community
Basic