R1 · evals

What this skill does

Build and run eval suites against agent prompts and skills.

Inputs

• Structured parameters typed per the skill contract.
• Credential handle supplied by the CloudSwarm vault when required.
• Mission context (budget, deadline, approval-path).

Outputs

• Structured result object the next skill can consume.
• Audit record committed to the run log.
• Cost and latency stamped on the step.

Example usage

• Composed inside a pre-built template; see the right column.
• Added to a custom agent via the builder: drop in, wire inputs, done.
• Scheduled or event-triggered via Core · scheduler.

Related templates

• Ship a React Native app end-to-end
• Inbound lead triage
• Proposal assembler
• Quote-to-contract
• Blog draft → publish
• Bug report → github issue