Skill Optimizer is a governance and maintenance tool designed to audit conversation history and local skill sets to improve triggering accuracy and workflow hygiene.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-optimizer-2
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 skill-optimizer-2 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 Skill Optimizer?
Skill Optimizer is a specialized utility for managing the lifecycle of Openclaw Skills within an AI agent environment. It acts as a technical auditor that parses current conversation history and local workspace directories to identify inefficiencies, such as missed triggers or overlapping capabilities. By analyzing how a user naturally interacts with the agent, it pinpoints exactly where skill metadata fails to match user intent.
Beyond simple discovery, the skill provides a structured framework for optimization. It transforms raw diagnostic data into a clear, actionable report, allowing developers to maintain a lean and high-performing agent environment. It ensures that every skill is properly documented, safe to use, and correctly synchronized with its underlying configuration files.
Skill Optimizer Use Cases
- Identifying why a specific skill failed to trigger during a complex conversation.
- Consolidating duplicate or near-identical skills that create architectural ambiguity.
- Refining weak skill descriptions and metadata to better align with natural user phrasing.
- Auditing local skill directories for stale configurations or risky actions lacking guardrails.
- Cleaning up local workspaces by identifying and deleting obsolete or redundant skill forks.
How Skill Optimizer Works
- The skill begins by inventorying all local skill roots, including hidden directories like .agents/skills and project-specific paths.
- It extracts evidence from the current thread, noting user goals, natural phrasing, and instances where skills were (or should have been) utilized.
- A diagnostic phase compares the thread evidence against the local inventory to identify issues like missed_trigger, weak_metadata, or duplicate_skill.
- It generates a comprehensive Audit Report following a strict schema, featuring an executive summary and evidence-based findings.
- The system presents a curated Action Queue where the user can choose to Fix, Merge, Delete, or Keep and Skip specific items.
- Only after explicit user confirmation does the skill perform file-system modifications such as rewriting metadata or merging files.
Skill Optimizer Setup
To integrate Skill Optimizer into your workspace, ensure it is placed within your local skills directory and that the reference schema is accessible.
# Navigate to your agent skills directory
cd ./.agents/skills
# Clone or move the skill-optimizer directory into place
# Ensure the references/report-schema.md file is present
ls ./skill-optimizer/references/report-schema.md
Skill Optimizer Data Schema & Taxonomy
Skill Optimizer utilizes a structured reporting format to ensure consistency. It organizes findings into the following taxonomy:
| Issue Type | Description |
|---|---|
| missed_trigger | Task matched a skill but the agent failed to invoke it. |
| weak_metadata | Descriptions miss common phrasing, leading to under-triggering. |
| duplicate_skill | Near-identical skill copies exist in the active environment. |
| overlap_skill | Multiple skills cover the same functional domain. |
| risky_skill | High-privilege actions found without sufficient safety confirmations. |
| stale_skill | Discrepancy between skill instructions and bundled source files. |
Skill Optimizer Advanced Features
- Evidence-based diagnostics using exact quotations from conversation logs for high-precision fixes.
- Multi-path inventory scanning that covers global, project-local, and hidden configuration directories.
- Non-destructive analysis mode that prevents any file modifications until a user selects a specific intervention.
- Heuristic-driven recommendations that prioritize safety and workflow stability over aggressive deletion.
- Standardized action menu to simplify the governance of complex Openclaw Skills ecosystems.
SKILL.md
Loading
Related Openclaw Skills
Li Lu China Opportunity Analysis
A professional analysis skill for identifying long-term investment opportunities and core assets within the context of China's massive modernization process.
Create PowerPoint (python-pptx)
A sophisticated automation skill for generating professional PowerPoint presentations programmatically using Python.
Web Profiler Bundle
A robust CLI toolkit for profiling, logging, and analyzing web request execution and performance metrics.
Popup CRO
Popup CRO is a specialized optimization skill designed to create high-converting overlays, modals, and banners that drive engagement without sacrificing user experience.
Knitify Health Chatbot
A research-powered health and science assistant that provides evidence-based answers with PubMed and ClinicalTrials.gov citations.
HWP Batch Convert
A powerful Windows-based automation tool for batch converting Hancom Office HWP/HWPX documents into various professional formats like PDF and DOCX.
Featured*
