Agent Skills: accelint-eval-architect

Use when users say "add an eval to this skill", "evaluate this skill/agent", "evaluate this tool/repo/pipeline", "how should I test this RAG bot / retrieval / parser", "set up DeepEval/Ragas", "should I use Promptfoo/Inspect", "is my eval any good", "audit my eval harness" — or when they describe symptoms like "my bot is hallucinating", "answers got worse after a prompt/model change", "this regressed and nothing caught it", or want CI/regression checks for prompts or LLM output. Assesses the target (an LLM skill OR a standalone tool repo, e.g. a RAG/retrieval pipeline), recommends DeepEval, Ragas, a deterministic harness, or human-review-only and scaffolds it; Promptfoo and Inspect AI are recommended with hand-off guidance when they fit (not scaffolded). Also audits existing eval harnesses for stale fixtures, toothless metrics, and uncalibrated thresholds. Make sure to use this whenever someone wants to measure, test, regression-check, or benchmark what a skill, agent, tool, or RAG pipeline produces.