ClaudeHemat for Openclaw

An intelligent model router that optimizes costs and performance by dynamically escalating tasks from Haiku to Sonnet or Opus based on complexity.

axellageraldinc
v1.0.3
Feb 16, 2026
0
1.3k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install claude-hemat

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install claude-hemat 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 ClaudeHemat?

ClaudeHemat is a specialized orchestration framework designed for the Openclaw Skills ecosystem. It functions as a traffic controller for large language model interactions, setting Claude Haiku 4.5 as the default engine for routine tasks to minimize latency and operational costs. By establishing a clear hierarchy of intelligence, it ensures that high-resource models are only utilized when strictly necessary.

The skill provides a structured approach to model selection within Openclaw Skills, preventing the waste of compute power on simple Q&A while maintaining the capability to handle extreme reasoning through automated session spawning. It bridges the gap between cost-efficiency and high-performance AI agent behavior.

ClaudeHemat Use Cases

  • Reducing operational costs by using Haiku 4.5 for simple file lookups and casual chat interactions.
  • Automatically escalating software development and architectural planning tasks to Sonnet 4.5.
  • Leveraging Opus 4.6 for deep research and mission-critical decision-making processes.
  • Standardizing model attribution across multi-agent workflows using Openclaw Skills.

How ClaudeHemat Works

  1. Receive user input and process it through the default Haiku 4.5 model for initial assessment.
  2. Apply predefined model rules to detect if the task involves coding, reasoning, or complex analysis.
  3. Execute a sessions_spawn command to initiate a specialized session with Sonnet 4.5 or Opus 4.6 if the complexity threshold is met.
  4. Process the complex request using the high-tier model and capture the output.
  5. Append the mandatory model attribution tag to the response and revert the primary active model back to Haiku 4.5.

ClaudeHemat Setup

To deploy ClaudeHemat within your Openclaw Skills environment, ensure your agent has permissions to manage multiple sessions and access the Anthropic API. Configure the model rules within your skill manifest as follows:

# Example of initiating a high-tier session within the skill logic
sessions_spawn(
  message: "Analyze this complex codebase",
  model: "anthropic/claude-sonnet-4-5",
  label: "Code Analysis"
)

ClaudeHemat Data Schema & Taxonomy

The skill maintains a strict metadata structure to track session state and model attribution within the Openclaw Skills framework.

Metadata Field Description Mandatory
default_model Set to anthropic/claude-haiku-4.5 Yes
session_label Short identifier for spawned tasks Yes
model_tag Explicit mention of the model used in the response Yes
escalation_trigger The logic rule that prompted the use of Sonnet or Opus Yes

ClaudeHemat Advanced Features

  • Conditional escalation logic that prevents Haiku from attempting reasoning or coding tasks.
  • Multi-session management allowing specialized models to handle concurrent complex sub-tasks.
  • Automatic session termination to prevent unnecessary resource consumption in Openclaw Skills.
  • Model-specific output formatting that ensures the user is always aware of the underlying intelligence tier being used.

SKILL.md


Loading

Related Openclaw Skills

Featured*