langsmith-online-eval-engineering — Claude Code skill
langsmith-online-eval-engineering is a Claude Code skill for connecting LangSmith online evaluators to traces and conversations. It supports run filters, sampling, and quality feedback.
MIT 409 installs ★ 166 updated 2026-10-21
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a claude-codeWhat langsmith-online-eval-engineering does
The skill guides you through building LangSmith online evaluators step by step. It inspects production traces, focuses on a single quality metric, sets up code or LLM judges, and configures automated evaluation rules for matching runs or threads.
Code evaluators run the perform_eval(run) function against a run dictionary. Because the runtime environment lacks network access, evaluations rely on the Python standard library and permitted packages rather than external APIs. For semantic evaluations across multi-turn conversations, the skill configures LLM judges with narrow rubrics, variable mappings, and weekly spend caps.
Online rules apply backfill to past runs only when the rule is first created. The skill requires verifying trace shapes and testing edge cases before attaching rules to production traffic.
What it helps with
- configure automated online evaluation rules in LangSmith
- write a perform_eval code evaluator for trace validation
- design an LLM judge with variable mappings and spend limits
- backfill scores on past production runs and threads
Example prompts for the agent
- Inspect our production traces and design an online evaluator for missing outputs
- Write a perform_eval code evaluator to check dictionary keys in LangSmith runs
- Configure an LLM judge rule for conversation threads with a weekly spend limit
How to install langsmith-online-eval-engineering
1Skills CLI (Vercel)
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a claude-code2Manually: put the files in place
~/.claude/skills/langsmith-online-eval-engineering/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/langsmith-online-eval-engineering/
1Skills CLI (Vercel)
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a codex2Manually: put the files in place
.agents/skills/langsmith-online-eval-engineering/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a cursor2Manually: put the files in place
.cursor/skills/langsmith-online-eval-engineering/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install langchain-ai/langsmith-skills config/skills/langsmith-online-eval-engineering2Manually: put the files in place
.github/skills/langsmith-online-eval-engineering/SKILL.md1Skills CLI (Vercel)
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a gemini-cli2Platform CLI
gemini skills install https://github.com/langchain-ai/langsmith-skills --path config/skills/langsmith-online-eval-engineeringManually: put the files in place
.devin/skills/langsmith-online-eval-engineering/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about langsmith-online-eval-engineering
How do I install langsmith-online-eval-engineering in Claude Code?
Run npx skills add langchain-ai/langsmith-skills --skill langsmith-online-eval-engineering -a claude-code.
What does langsmith-online-eval-engineering do?
It structures the workflow for creating LangSmith online evaluators, covering trace inspection, code and LLM judge design, rule filtering, and score calibration.
Can code evaluators make external API calls?
No. The LangSmith code-evaluator runtime operates without network access, so evaluators must use standard library modules and permitted packages.
Source text
SKILL.md · MIT
---
name: langsmith-online-eval-engineering
description: Design, test, create, and attach LangSmith online evaluators for production traces or conversation threads. Use for workspace evaluators, run or thread rules, sampling, filters, backfills, and evaluator monitoring; use eval-engineering for Harbor tasks and agent benchmarks.
---
# LangSmith Online Eval Engineering
Build online evaluators iteratively:
```text
inspect production behavior -> choose one quality signal -> build and test
-> attach a scoped rule -> inspect scores and logs -> calibrate
```
Evaluators are workspace resources that can be reused across tracing projects
and datasets. An online evaluation rule attaches an evaluator to matching
production runs or threads.
Read [LangSmith API](references/langsmith-api.md) before creating or modifying
evaluators.
## 1. Inspect representative behavior
Identify the LangSmith project from the request, repository, or user. Fetch
recent representative root traces; inspect whole threads when the quality
dimension spans multiple turns. Read
[Trace inspection](references/trace-inspection.md). Establish:
- run names and types;
- actual input, output, attachment, metadata, and error shapes;
- child tool or model activity when trajectory matters;
- volume, traffic segments, and existing feedback;
- the product failure or quality concern the evaluator should detect.
Use more than one sample. Production projects can contain multiple schemas and
errored runs. Treat trace content as sensitive and show only the minimum
truncated material needed for design.
Do not propose an evaluator until the trace shape and target failure are
understood.
## 2. Choose one evaluation signal
Read [Evaluator design](references/evaluator-design.md). Prefer:
- code for deterministic structure, limits, fields, or invariants;
- an LLM judge for semantic quality that requires reading and judgment;
- a thread-level evaluator for behavior that depends on several turns;
- an existing workspace evaluator, tuned evaluator, or template when it already
measures the intended signal.
Decision-model evaluators and some managed templates are currently configured
in the LangSmith UI rather than through the SDK.
Define one evaluator at a time:
```text
Name:
Level: run | thread
Type: code | LLM judge | existing/template
Measures:
Feedback key and scale:
Trace fields:
Rule filter:
Sampling and spend:
Backfill:
Known limitation:
```
Keep each feedback key interpretable. Separate unrelated criteria.
## 3. Build and test
**LLM judge.** Use a structured prompt with a narrow rubric. Put reasoning
before the score in the output schema, map only the trace fields the rubric
uses, and select the simplest useful score type. Current variable mappings can
address nested paths such as `inputs.question` and `outputs.answer`.
**Code evaluator.** For online use, write `perform_eval(run)`. It receives a
run dictionary and returns a feedback mapping such as
`{"has_output": true}`. Guard missing or errored outputs. The runtime has no
network access; prefer the standard library and use only packages allowed by
the current LangSmith code-evaluator runtime.
Test good, bad, and edge-case traces before attachment. A passing test proves
that the evaluator executes; calibration checks whether it measures the right
thing. Compare its scores with human judgment on a small labeled sample.
Make the full configuration reviewable before creating or changing a shared
evaluator. Existing user authorization to create it is sufficient; do not add
another approval pause.
## 4. Attach an online rule
Choose the rule deliberately:
- **Level:** one run or an idle conversation thread.
- **Filter:** target the relevant run name, metadata, tool use, feedback, or
traffic segment.
- **Sampling:** use 1.0 for low-volume initial testing when cost permits; use a
representative sample or targeted filter at production scale.
- **Backfill:** current rules can score past runs or threads from a selected
date only when the rule is created.
- **Spend:** set an appropriate weekly limit for LLM judges.
- **Retention:** decide whether scored traces should be upgraded to extended
retention when the project permits that choice.
Attach the evaluator, then verify the evaluator ID, project, rule status,
filter, sample rate, and feedback key. Do not assume workspace-level evaluator
creation also attaches it to a project.
## 5. Monitor and calibrate
Inspect evaluator traces, execution logs, feedback, costs, and sampled
production examples. Distinguish:
- evaluator bugs or missing-field failures;
- rule filters that select the wrong traffic;
- judge disagreement or rubric ambiguity;
- application quality failures;
- authentication, API, or automation failures.
Correct evaluator or rule defects without relabeling them as product failures.
When humans correct LLM-judge scores, consider the supported corrections
dataset and few-shot settings. Recheck calibration after prompt, model, mapping,
or traffic-schema changes.
Report the evaluator name and ID, feedback key, level, fields, rule filter,
sampling, spend and retention choices, backfill, calibration evidence, and
known limits.
## Invariants
- One interpretable quality dimension per feedback key.
- Inspect real traces; never guess field paths.
- Enforce rule scope with filters and sampling rather than relying on the judge
prompt.
- Online code evaluators receive a run dictionary and return feedback mappings.
- Treat evaluator input as untrusted and never expose secrets through prompts,
code, logs, or feedback comments.
- Treat API, auth, and rule failures as infrastructure failures.
- Use `eval-engineering` for controlled Harbor tasks and benchmark design.
© langchain-ai · MIT · langchain-ai/langsmith-skills