recursive-language-model
RLM (Recursive Language Model) Skill
This skill enables processing of arbitrarily long documents by treating them as an external environment and recursively calling sub-LLMs over chunks of the content.
Overview
The RLM pattern follows the architecture from the paper:
- Root Agent: Main agent orchestrating the task
- Sub-LLM (llm_query): Subordinate agent for chunk-level analysis
- External Environment: Persistent Python REPL with document state
Requirements
- Python 3.8+
- An agentic harness with code execution and subordinate agent capabilities
File Structure
More from viktor-ferenczi/skills
python-guidelines
Guiding principles for writing clear, concise, human readable and maintainable Python code.
20gcp-batch-inference
Running batch inference on Google Cloud (also known as Vertex AI)
18silent-cli
Environment variables and parameters for running command line programs reliably in non-interactive environments (unattended). Includes silent modes, color/disable TTY, and reduced output options for 155 CLI tools.
12busybox-on-windows
How to use a Win32 build of BusyBox to run many of the standard UNIX command line tools on Windows.
11stabilization-loop
Stabilizes a software project by repeatedly running and testing it in a loop, fixing any issues.
6consistency-check
Checks the internal consistency of a software project, fixes any issues found.
3