Control Center

The Control Center is your primary dashboard for understanding how you use Claude. It shows live data about conversations, skill usage, and performance metrics. Combined with the MCP analytics tools, it provides the feedback loop you need to keep your skill library effective.

Conversations

The conversation feed shows every conversation tracked through your connection. Each entry includes:

• Status: Whether the conversation was successful, needs review, or had issues
• Skills used: Which skills Claude applied during the conversation
• Session analysis: AI-generated quality assessment of the conversation
• Timestamps: When the conversation happened and how long it lasted

You can filter conversations by status (all, successful, needs review, failed) to quickly find conversations that need attention.

Usage metrics

The Control Center displays key performance metrics:

Metric	What it measures
APM (Actions Per Minute)	How actively you are using Claude
EAPM (Effective Actions Per Minute)	Actions that resulted in useful output
Concurrency	How many simultaneous conversations you run
Throughput	Total volume of work processed through Claude
Health score	An overall 0-100 score assessing the health of your skill usage

Skill effectiveness

The most actionable metric is skill effectiveness. This tells you which skills are producing good results and which are not. A skill with high usage but low effectiveness needs improvement. A skill with zero usage might need better naming or should be retired.

MCP analytics tools

Beyond the visual dashboard, the Skill Manager MCP server provides two analytics tools you can use directly in conversation with Claude.

analyze_usage

Ask Claude:

"Analyse my overall usage and give me a health report."

Claude will use the analyze_usage tool to generate a comprehensive report including:

• Health score (0-100) with assessment: Excellent, Good, Fair, or Needs Attention
• Today's summary: What happened today, which skills were used, any issues
• Last 7 days: Daily summaries showing trends and patterns
• Skill effectiveness rankings: Your best and worst performing skills
• Entity usage summary: How often each skill, agent, and tool is used
• Unused skills: Skills that exist but have not been used recently
• Performance metrics: APM, success correlation, and throughput data

This is the single most useful command for understanding whether your skill library is working.

analyze_skill_usage

For deeper analysis of a specific skill, ask Claude:

"Analyse the usage of my Brand Voice skill and suggest improvements."

Claude will use the analyze_skill_usage tool to provide:

• Usage frequency: How often the skill is invoked
• Effectiveness data: Whether the skill produces good results
• Co-usage patterns: Which other skills are commonly used alongside this one
• Session analyses: Detailed breakdowns of conversations where the skill was used
• AI-generated improvement suggestions: Specific changes to make the skill more effective

This is how you identify exactly what to change in a skill to improve it.

The improvement workflow

Analytics without action is just data. Here is a practical workflow for using analytics to improve your skills.

Weekly review (15 minutes)

1. Check your health score. If it is below 70, you have significant room for improvement. Above 85 is excellent.

2. Review unused skills. Ask Claude: "Which of my skills have not been used in the last two weeks?" Consider whether unused skills need better names, need to be promoted to your team, or should be retired.

3. Check the bottom performers. Ask Claude: "Which of my skills have the lowest effectiveness scores?" Pick the worst one and analyse it in detail.

4. Update one skill. Based on your analysis, update one skill per week. Small, focused improvements compound over time.

After a bad conversation

When a conversation produces poor results:

1. Find the conversation in the Control Center
2. Check which skills were applied
3. Determine whether the skill instructions were followed (a skill problem) or whether the wrong skill was matched (a naming/description problem)
4. Edit the skill to address the issue

Before adding new skills

Before creating a new skill, check your analytics:

1. Are any existing skills covering similar ground? Update rather than duplicate.
2. What are your most-used skills? New skills in the same domain will likely get used.
3. What tasks do you perform that have no skill coverage? These are your gaps.

Gamification and progress

systemprompt.io tracks your progress through a gamification system. As you use skills, create new ones, and improve existing ones, you earn XP and advance through ranks.

Your current rank, XP, and progress are visible in the top-right corner of every admin page. The achievement system rewards specific milestones like creating your first skill, reaching 100 conversations, or maintaining a health score above 80.

The gamification system encourages the behaviours that make the platform more valuable: regular usage, skill refinement, and thoughtful organisation.

Profile and insights

The Profile & Insights page provides a deeper view of your usage patterns over time:

• Your conversation history and trends
• Skill usage distribution
• Quality trends over time
• AI-generated profile reports that summarise your usage patterns and suggest improvements

Use this page for monthly reviews of your overall Claude usage strategy.

Reporting

The Control Center includes a reporting feature. Click the Report tab to generate a comprehensive report of your usage. Reports include:

• Conversation summaries
• Skill effectiveness data
• Health metrics
• Improvement recommendations

These reports are useful for sharing with your team or manager to demonstrate the value of your Claude usage.

Troubleshooting

No conversations appearing

• Verify your connection is active.
• Start a new conversation — the Control Center tracks conversations initiated after the connection was established.
• Check that the Skill Manager connector is connected, not just the marketplace.

Health score seems inaccurate

• The health score needs at least a week of data to be meaningful.
• Check for conversations that were marked as "needs review" — these lower the score.
• Use analyze_usage for a breakdown of what is affecting the score.

Analytics tools not responding

• Verify the Skill Manager connector is active.
• Start a new conversation and try again.
• Check your network connection to systemprompt.io.

Next steps

• Editing Skills — improve skills based on your analytics
• Adding and Personalising Plugins — expand your skill library
• Creating Skills — build new skills for gaps you identify