Evaluate LLM extraction pipeline changes against a baseline and produce merge, regression, or defer verdicts.
npx skills add https://github.com/idank/explainshell --skill eval-llm
Please help me install this Agent Skill. Skill: eval-llm Source: https://github.com/idank/explainshell/tree/main/.claude/skills/eval-llm Command: npx skills add https://github.com/idank/explainshell --skill eval-llm
Evaluates changes to an LLM-based extraction workflow end-to-end against a clean baseline to determine safety and impact before committing.
Invoke the evaluation workflow to run a baseline and candidate, then compare results to classify the change and decide on commit.
High-intent search queries and answers about installing and using this skill.
To validate LLM prompt changes, run a baseline and candidate evaluation to compare extraction results and produce a structured verdict (merge, regression, or defer) with per-page diagnostics.
End-to-end LLM evaluation tests changes to prompt, chunking, or post-processing code against a clean baseline to measure safety and impact before committing updates.
Yes, LLM evaluation supports validating chunking changes and post-processing updates across extraction components, applying baseline and candidate runs to classify the impact.
Invoke the evaluation workflow to run both baseline and candidate LLM extraction runs, compare the results, and generate a verdict classifying whether to merge, regress, or defer.
LLM evaluation produces structured verdicts of merge, regression, or defer for prompt tweaks, providing per-page diagnostics to determine safety for commit.
Yes, a clean baseline is required to benchmark LLM extraction changes against the candidate version, ensuring accurate comparison and reproducible evaluation results.
ResearAIStandardize a reference object for quest evaluation with plan and checklist requirements.
prestariusAudit Claude Code skills for overlap, freshness, and trigger integrity.
MisLinkDraft SKILL.md files, generate test prompts, and run with-skill versus baseline evaluations.
masermediagroup-stackEvaluate LLM systems with automated metrics, LLM-as-judge rubrics, and standardized benchmarks.
