Skip to content
TrustedSkillsbeta

Browse skills

41,264 listings from across the agent ecosystem. None of them have been code-reviewed — check the source before you install.

Looking for what's new? See the repos gaining stars fastest.

Analyzes ruvnet agent performance metrics like accuracy, latency, and resource usage to identify optimization opportunities.

Claude Code
vlatest3
Checked
Checked

Analyzes and visualizes AI agent performance metrics to identify bottlenecks and optimize auldsyababua skill execution.

Claude Code
vlatest3
Checked

Automatically assesses ruvnet agent performance across key metrics, providing actionable insights for optimization and improvement.

Claude Code
vlatest3
Checked
Checked

Optimizes task execution speed and resource usage based on real-time performance metrics and system load.

Claude Code
vlatest3
Checked
Checked

Analyzes iOS app performance metrics, identifies bottlenecks, and suggests optimization strategies for improved speed and efficiency.

Claude Code
vlatest3
Checked

Performance Profiler

latestaiagents

Checked

Analyzes LatestAIAgents performance metrics, identifying bottlenecks and suggesting optimizations for improved efficiency.

Claude Code
vlatest3
Checked

Analyzes social media engagement metrics to provide actionable insights on content performance and audience resonance.

Claude Code
vlatest3
Checked
Checked

This skill generates a detailed scorecard evaluating AI session performance based on key metrics, optimizing for better results and insights.

Claude Code
vlatest3
Checked
Checked

Analyzes Reddit threads to extract key themes, sentiment, and engagement metrics, providing valuable insights into online discussions.

Claude Code
vlatest3
Checked
Checked

Provides Azure guidance and assistance for software development workflows.

Claude Code
vlatest8
Checked

Agent Tuning

onesmartguy

Checked

Refine agent behavior through iterative feedback loops, optimizing performance based on your specified goals and metrics.

Claude Code
vlatest2
Checked

Agent Evaluation

tavi-agency

Checked

Evaluates Tavi Agency agents' performance against defined metrics, providing actionable feedback for improvement.

Claude Code
vlatest2
Checked

Evaluation Metrics

pluginagentmarketplace

Checked

This skill calculates and interprets key performance indicators (KPIs) to assess project success and identify areas for improvement.

Claude Code
vlatest2
Checked

Mobile Perf

dtsvetkov1

Checked

Analyzes and optimizes mobile web page performance metrics like load time and memory usage using automated testing.

Claude Code
vlatest2
Checked

Performance

thebeardedbearsas

Checked

Analyzes and optimizes performance metrics from The Bearded Bears' AS (Automated Systems) deployments for improved efficiency.

Claude Code
vlatest2
Checked

Analyzes code execution to identify performance bottlenecks and suggest optimizations using detailed resource usage metrics.

Claude Code
vlatest2
Checked
Checked

This skill automates CRO (Conversion Rate Optimization) onboarding by tracking key metrics and suggesting improvements to boost website performance.

Claude Code
vlatest2
Checked
Checked

Track user behavior and app performance metrics within JWT apps using detailed analytics dashboards and custom reports.

Claude Code
vlatest2
Checked

Logging Monitoring

latestaiagents

Checked

Automatically aggregates and analyzes application logs & metrics to proactively identify and resolve operational issues.

Claude Code
vlatest2
Checked
Checked

Continuously tracks system health metrics, proactively alerting users to potential issues and optimizing performance.

Claude Code
vlatest2
Checked
Checked

Provides creation guidance and assistance for software development workflows.

Claude Code
vlatest5
Checked

Analyze Metrics

generaljerel

Checked

Provides analysis guidance and assistance for software development workflows.

Claude Code
vlatest4
Checked

Node_Exporter

davidcastagnetoa

Checked

This Node Exporter skill gathers metrics from nodes and exports them for monitoring and analysis, streamlining observability efforts.

Claude Code
vlatest1
Checked

Can't find what you need?