Evaluate Claude Code sessions with configurable graders and run pipelines.
npx skills add https://github.com/Clean1ines/crm_bot --skill eval-harness-clean1ines
Please help me install this Agent Skill. Skill: eval-harness Source: https://github.com/Clean1ines/crm_bot/tree/main/.agents/skills/ecc/eval-harness Command: npx skills add https://github.com/Clean1ines/crm_bot --skill eval-harness-clean1ines
Formal evaluation framework to systematically verify Claude Code sessions against defined criteria, enabling predictable and auditable improvements.
Create an initial eval definition and run the first capability/regression checks to bootstrap the workflow.
High-intent search queries and answers about installing and using this skill.
Setting up regression testing for AI agent sessions requires configuring evaluation definitions and run pipelines to systematically verify behavior against defined criteria. You can track reliable agent behavior across model versions using repeatable test suites.
Eval-driven development is a formal evaluation framework to systematically verify AI agent sessions against defined criteria, enabling predictable and auditable improvements. It standardizes capability and regression evaluation objectives.
Automated grading for AI capability evaluations supports deterministic code-based graders and model-based graders. This flexible assessment approach allows repeatable test suites without requiring manual human review for every check.
Configuring a capability evaluation pipeline requires defining evaluation definitions, graders, and run pipelines. You bootstrap the workflow by creating an initial eval definition and running the first capability checks.
The evaluation framework supports human review alongside deterministic code-based and model-based graders for flexible assessment. This enables commissioning audits and regression tracking across mixed grading strategies.
Formal regression tracking for AI models is needed when verifying reliable agent behavior across model versions. It applies to workflows requiring capability evaluations and repeatable test suites to ensure predictable improvements.
JamkrisAutomate eval-driven testing workflows for Gemini CLI sessions with deterministic graders.
cskiroAutomate Playwright end-to-end test setup, generation, and regression detection.
KunCheng-HeCreate, refine, and benchmark custom AI skills with Python-based test cases.
ysyecustCodify and automate Claude Code session evaluations with pass@k metrics.
