Loading
Define measurable success criteria for an LLM application and design the evaluations that test them. Produces an approved plan covering criteria, eval methods, graders, and test-case coverage. Optionally generates an eval harness and runs it, each behind its own human approval gate.
npx skills add Reactive-Skills/skills --skill llm-eval-designnpx -y @reactive-skills/axi invoke llm-eval-design16
Hierarchical nodes
INIT
Bootstrap entrypoint
0
Whitelisted for safety
5
Automated projections
Visual state machine topology parsed directly from the official STATECHART.md in Reactive-Skills/skills:
Each state is an isolated operational slice. Transitions occur only when typed signals satisfy programmatic guard conditions.
Operational state INIT
Outgoing Transitions
Operational state ACTIVE
Outgoing Transitions
Operational state ACTIVE.INTAKE
Outgoing Transitions
Operational state ACTIVE.CRITERIA
Outgoing Transitions
Operational state ACTIVE.EVAL_DESIGN
Outgoing Transitions
Operational state ACTIVE.GRADING
Outgoing Transitions
Operational state ACTIVE.APPROVE_PLAN
Outgoing Transitions
Operational state ACTIVE.HARNESS
Outgoing Transitions
Operational state ACTIVE.APPROVE_RUN
Outgoing Transitions
Operational state ACTIVE.RUN
Operational state ACTIVE.RUN.CALIBRATE
Outgoing Transitions
Operational state ACTIVE.RUN.SCORE
Outgoing Transitions
Operational state ACTIVE.FINALIZE
Outgoing Transitions
Operational state COMPLETE
Operational state BLOCKED
Operational state BYPASS_DETECTED
Inspect or drive this state machine directly via the token-efficient AXI CLI:
Check state:
Inspect hierarchy:
Tail events:
Read models automatically rendered upon transition:
Add to your Claude Desktop, Cursor, or VS Code settings:
{
"mcpServers": {
"reactive-skills": {
"command": "npx",
"args": [
"-y",
"@reactive-skills/axi",
"mcp"
]
}
}
}