Research: langsmith-evaluator
Cached research evidence for langsmith-evaluator (not authority).
Purpose
Section titled “Purpose”Official from langchain-ai/langsmith-skills (early dev, 131 stars). Create custom evaluators (with helpers). Requires LANGSMITH_API_KEY + cloud access per README. Part of observability for LLM apps. Complements langchain-skills.
Harness Coverage
Section titled “Harness Coverage”antigravity,claude-code,… group targets.
Trust And Risks
Section titled “Trust And Risks”trust_tier=needs-inspection; status=inspect-then-install; requires secret API key (user provisioned); early dev; scripts included may exec. Official LangChain but scoped to LangSmith tracing/eval.
Install Prerequisites
Section titled “Install Prerequisites”npx skills add langchain-ai/langsmith-skills —skill … -y -g -a [..]; set LANGSMITH_API_KEY; status=inspect-then-install.
Upstream Maintainer
Section titled “Upstream Maintainer”Comparable Alternatives
Section titled “Comparable Alternatives”Other trace/eval skills (e.g. langfuse, phoenix).
> Evidence from public web sources, GitHub (official orgs 2026), curated catalog; research evidence, not authority. External audit advised.