A specialized controller skill for managing massive inputs through external context slicing and recursive sub-agent orchestration.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install rlm-controller
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install rlm-controller using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is RLM Controller?
The RLM Controller provides a robust, policy-driven framework designed to handle inputs that exceed the standard context windows of modern language models. As part of the wider ecosystem of Openclaw Skills, it treats large datasets, repository-scale codebases, and dense logs as external context files. By peeking, searching, and chunking these files, it allows agents to navigate vast amounts of information without losing technical precision.
This skill is particularly effective for workflows that require deep, multi-pass analysis. It utilizes a structured approach to spawn sub-agents that handle specific data slices, ensuring that the primary agent remains focused on high-level orchestration while the heavy lifting of data processing is distributed safely and efficiently.
RLM Controller Use Cases
- Analyzing repository-scale codebases for security vulnerabilities or architectural patterns
- Processing massive server logs and datasets that exceed token limits
- Performing recursive sub-tasks on extremely long legal or technical documentation
- Orchestrating complex, multi-agent analysis workflows across large file systems
How RLM Controller Works
- Store the large input file as an external context artifact using the context storage script.
- Use the automated planner to identify specific keywords or slices that require deeper analysis.
- Organize identified tasks into asynchronous batches to optimize processing time.
- Spawn sub-agent sessions to handle individual data slices independently.
- Aggregate the structured results from all sub-agents back into the root session for final synthesis.
RLM Controller Setup
To begin using this skill within your Openclaw Skills environment, ensure the helper scripts are accessible. You can store your first input for analysis using the following command:
python3 scripts/rlm_ctx.py store --input large_dataset.jsonl
After processing, you can purge temporary artifacts and maintain your environment with the included cleanup utility:
./scripts/cleanup.sh
RLM Controller Data Schema & Taxonomy
| Artifact Type | Description |
|---|---|
| Context Storage | External files used to hold the primary input data outside the LLM window |
| Planning Manifests | JSONL files that define how the input is sliced and which tools are invoked |
| Policy Documentation | Markdown files defining safety limits, recursion depth, and call thresholds |
| Execution Traces | Log summaries used to audit the lifecycle of sub-agent calls |
RLM Controller Advanced Features
- Strict recursion depth limits to prevent infinite sub-agent loops and control costs
- Integrated secret redaction to automatically scrub sensitive patterns from data before processing
- Non-blocking asynchronous session spawning for high-throughput data analysis
- Customizable cleanup rules to manage disk space and artifact retention policy
SKILL.md
Loading
Related Openclaw Skills
Serper.dev Google Search
Perform programmatic Google searches via the Serper.dev API to provide your AI agents with structured, real-time web data without the need for manual scraping.
Sales Analyzer
An AI-powered sales coach that analyzes meeting transcripts to diagnose sales techniques, map client relationships, and provide actionable improvement plans.
Get Notes Sync (get-biji)
A powerful synchronization tool that automates the export of AI-transcribed voice notes from biji.com into organized local Markdown files.
Daily Scanner: AI Note Inventory & CRM Automation
An automated AI agent skill that scans, classifies, and syncs voice notes to CRMs and task managers daily.
Monolith — Crypto Wallet
A hardware-secured crypto wallet for AI agents utilizing Apple Secure Enclave and smart contract account abstraction for safe on-chain transactions.
Looplink Skill
The Looplink skill enables AI agents to seamlessly interact with the looplink.app platform for social content organization and bookmarking.
METADATA
Featured*
