| name | cekura-coordinator |
| description | Use when the user asks "what can Cekura do", "what commands are available", "help me with Cekura", "what skills do I have", "show me Cekura features", "what's available", "how do I use Cekura", or needs guidance on which Cekura skill to use for their task. Also relevant as the entry point when a user has just installed cekura-skills for the first time.
|
| license | MIT |
| compatibility | Requires a Cekura account (https://dashboard.cekura.ai) — sign in via OAuth or use an API key. |
| metadata | {"author":"cekura","version":"0.3.0"} |
Cekura Coordinator
Purpose
Route users to the right Cekura skill or command based on what they need. This is the "front desk" — it knows everything available across all Cekura plugins and helps users find the right tool.
Performing Platform Actions
When this skill suggests creating, listing, updating, or evaluating something on Cekura, prefer using available platform tools over describing API calls or dashboard steps. In Claude Code with the Cekura plugin installed, these tools are auto-configured and handle authentication, parameter validation, and error handling for you. Fall back to direct API endpoints or dashboard guidance only when no tools are available in the current session.
When This Skill Loads
- User just installed cekura-skills and asks "what can I do?"
- User asks for help or doesn't know which command to use
- User describes a task and you need to route them
Available Skills & Commands
Skills activate when the user describes a relevant task. Commands are slash commands available only in Claude Code with the plugin installed.
Skills
| Skill | Use when |
|---|
cekura-onboarding | New to Cekura, first-time setup, platform walkthrough |
cekura-create-agent | Set up a voice AI agent — provider, mock tools, KB, dynamic vars |
cekura-self-improving-agent | Auto-tune agent prompts from eval results — diagnose → propose → apply → re-validate |
cekura-metric-design | Design new metrics, improve existing ones, metric best practices |
cekura-metric-improvement | Improve metric accuracy through feedback cycle (labs workflow) |
cekura-predefined-metrics | Catalog of built-in metrics — what each does, costs, constraints, configuration |
cekura-eval-design | Design evaluators, test suites, coverage strategy, conditional actions |
Commands
| Command | Use when |
|---|
/cekura-onboarding | Run the guided onboarding flow (state-aware, picks up where you left off) |
/setup-mcp | Configure the Cekura MCP server (Claude Code only) |
/upgrade-skills | Update all Cekura skills to the latest version |
/report-bug | Report a bug — files a GitHub issue, optionally attempts a fix |
/create-metric | Create or update a metric via API |
/list-metrics | List metrics for an agent or project |
/evaluate-calls | Run metrics on specific calls |
/improve-metric | Full improvement cycle: collect feedback, run labs, auto-improve |
/manual-create-update-eval | Create or update a single evaluator with full field walkthrough |
/autogen-eval | Auto-generate evaluators (or bulk create from CSV/JSON) |
/list-evals | List evaluators for an agent or project |
/run-evals | Execute evaluators (run test scenarios) |
/eval-results | Check results from a test run |
/cekura-report | Full end-to-end quality report — generates 10 evals, runs them, produces structured analysis |
Routing Guide
When the user describes what they need, route them:
| User Need | Route To |
|---|
| "I'm new to Cekura" / first-time setup | cekura-onboarding skill |
| "Set up my agent" / "connect my voice agent" | cekura-create-agent skill |
| "Configure MCP" / "MCP not working" | /setup-mcp command |
| "Create metrics for my agent" | cekura-metric-design skill |
| "My metrics are giving wrong results" | /improve-metric command (or cekura-metric-improvement skill for full cycle) |
| "I need to test my agent" | cekura-eval-design skill |
| "Generate test scenarios" | /autogen-eval command |
| "Create a specific test scenario" | /manual-create-update-eval command |
| "Run my tests" | /run-evals command |
| "Check test results" | /eval-results command |
| "Create a metric that checks X" | /create-metric command (or cekura-metric-design skill for complex metrics) |
| "Update this metric" | /create-metric command (handles both create and update) |
| "Evaluate calls against metrics" | /evaluate-calls command |
| "Update my skills" | /upgrade-skills command |
| "What metrics should I have?" | cekura-metric-design skill (baseline metrics section) |
| "Help me improve this metric" | /improve-metric command |
| "Leave feedback on a metric result" | /improve-metric command (Phase 1: feedback collection) |
| "Set up production monitoring" | cekura-onboarding skill (Phase 6) + observability docs |
| "Add mock tools" / "set up tools" | cekura-create-agent skill (Phase 4) |
| "Upload knowledge base" | cekura-create-agent skill (Phase 5) |
| "Something's broken" / "file a bug" | /report-bug command |
| "Improve my agent" / "auto-tune from eval results" | cekura-self-improving-agent skill |
| "Which built-in metrics are available?" / "what does Hallucination Detection cost?" | cekura-predefined-metrics skill |
| "Fix this prod call bug" / "reproduce and test a fix" | cekura-self-improving-agent skill |
| "Run a full quality report" / "generate evals and run them end-to-end" | /cekura-report command |
Typical User Journeys
Journey 1: Brand New User
/setup-mcp → Configure MCP server for API access (Claude Code plugin users)
- cekura-onboarding → Set up account and project
- cekura-create-agent → Add agent with provider, mock tools, KB, dynamic vars
- cekura-onboarding → Enable pre-defined metrics, generate first evaluators
- cekura-onboarding → Run first tests, review results
- cekura-metric-design → Create custom metrics based on what they learned
- cekura-eval-design → Build targeted test suites
Journey 2: Has Agent, Needs Testing
- cekura-eval-design → Design test suite
/autogen-eval → Auto-generate evaluators
/run-evals → Execute tests
/eval-results → Review results
Journey 3: Has Metrics, Needs Improvement
/improve-metric → Full cycle: collect feedback, check readiness, auto-improve
/evaluate-calls → Validate changes
Journey 4: Production Monitoring
- cekura-metric-design → Design observability metrics
/create-metric → Deploy metrics (create or update)
/evaluate-calls → Validate on sample calls
API Access
For Claude Code plugin users: each plugin auto-configures access to the Cekura API. If commands or platform operations aren't working, run /setup-mcp to configure the connection.
For other clients (Cursor, Codex, npx skills installs, etc.): use the Cekura dashboard at https://dashboard.cekura.ai or call the API directly using your API key.
Next Steps
This skill routes — it doesn't perform tasks itself. After confirming the user's need, invoke the matching skill:
- New to Cekura → cekura-onboarding
- Connecting an agent → cekura-create-agent
- Auto-tuning an agent prompt from eval results → cekura-self-improving-agent
- Designing metrics → cekura-metric-design
- Improving metric accuracy → cekura-metric-improvement
- Picking which built-in metrics to use → cekura-predefined-metrics
- Designing test scenarios → cekura-eval-design
- Fixing a production call bug end-to-end → cekura-self-improving-agent
Documentation