A zero-dependency efficiency auditor that provides Lighthouse-style grading and performance optimization for Openclaw Skills instances.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawzembic
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 clawzembic 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 Clawzembic?
Clawzembic is a specialized diagnostic and optimization tool designed to ensure your Openclaw Skills environment remains lean and cost-effective. Functioning like a Lighthouse audit for AI agents, it scans your installation to identify "bloat"—such as oversized context files, inefficient cron configurations, and stale sessions that consume unnecessary tokens. By providing a clear letter grade from A+ to F, it helps developers maintain high performance and lower API costs.
Built with a zero-dependency philosophy, this skill relies strictly on the Python standard library, making it extremely portable. Whether you are managing a local setup or a complex remote instance, Clawzembic offers the transparency needed to right-size your models and keep your Openclaw Skills workspace running at peak efficiency.
Clawzembic Use Cases
- Reducing token consumption by identifying oversized MEMORY.md and workspace files.
- Auditing background cron jobs to prevent main-session pollution and model over-spending.
- Cleaning up stale sessions that cause context window bloat and increased latency.
- Benchmarking the health of remote Openclaw Skills instances via SSH.
- Automating performance remediation with guided fix suggestions.
How Clawzembic Works
- The tool initiates a comprehensive scan of the local .openclaw directory or connects to a remote host via SSH.
- It analyzes six critical efficiency categories: context injection, cron health, session bloat, config integrity, skill footprint, and transcript size.
- Weights are applied to each category (e.g., 25% for Context Injection and Cron Health) to generate a numeric score and corresponding letter grade.
- Detailed findings are generated, highlighting specific files or configurations that exceed efficiency thresholds.
- If the --fix flag is used, the skill suggests or applies automated remediations to bring the Openclaw Skills instance back to optimal health.
Clawzembic Setup
Clawzembic is ready to use with Python 3.8+ and has no external dependencies. Use the following commands to audit your Openclaw Skills setup:
# Audit the local instance
bash skills/clawzembic/lean-audit.sh
# Audit a remote instance via SSH
bash skills/clawzembic/lean-audit.sh --remote user@host
# View automated fix suggestions for an underperforming instance
bash skills/clawzembic/lean-audit.sh --fix
Clawzembic Data Schema & Taxonomy
Clawzembic organizes its audit data across six performance pillars. The report identifies specific metrics that impact the health of Openclaw Skills:
| Pillar | Metric | Impact |
|---|---|---|
| Context Injection | File size of MEMORY.md and workspaces | Token Waste |
| Cron Health | Model selection and session isolation | Operational Cost |
| Session Bloat | History length and waste ratio | Inference Latency |
| Config Health | Heartbeat frequency and compaction settings | System Overhead |
| Skill Bloat | Number of skills injected into system prompt | Prompt Efficiency |
| Transcript Size | Log and file disk usage | Storage Management |
Clawzembic Advanced Features
- Zero-dependency architecture using only the Python standard library for maximum compatibility.
- Remote audit capabilities allowing for agent-less diagnostics of external Openclaw Skills deployments.
- JSON output flag (--json) for integrating audit results into monitoring dashboards or CI/CD pipelines.
- Automated remediation mode to quickly fix common configuration errors and reduce session bloat.
- Support for custom directory paths to analyze non-standard installations of Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Gold Price Checker
An automated financial tracking tool providing real-time data on domestic and international gold markets, retail jewelry, and investment bars.
China Scenic Spots Top 100
A comprehensive data-driven skill for AI agents to provide detailed information, recommendations, and itinerary planning for China's top 100 tourist attractions.
Anime Calendar
An automated anime tracking tool that provides weekly update schedules and seasonal release information from major platforms.
AWS Terraform Security Reviewer
A security-focused AI agent skill designed to audit AWS Terraform configurations and plan files for misconfigurations before deployment.
twitter-openclaw
A comprehensive interface for AI agents to interact with Twitter/X, enabling tweet reading, posting, engagement, and timeline management.
Home Assistant Skill
A comprehensive integration allowing AI agents to control, monitor, and automate over 25 domains of smart home devices through the Home Assistant REST API.
Featured*
