Editorial illustration for LangSmith CLI adds three portable skills for coding agents in the repo
LangSmith CLI Boosts Coding Agents with 3 New Skills
LangSmith CLI adds three portable skills for coding agents in the repo
A coding agent is only as good as its environment. Give it the right tools, and it transforms. LangSmith CLI just gave agents three new portable skills, trace, dataset, and evaluator, that snap directly into any repo.
These aren't abstract concepts. They are the operational backbone of AI engineering: tracing what your code does, building datasets from real interactions, and evaluating agents against them. The result?
Claude Code's performance on LangSmith tasks jumped significantly. That’s the power of portability. That’s what happens when you stop asking agents to guess and start equipping them with purpose-built skills.
We’re sharing a set of LangSmith skills that can be ported to any coding agent that supports skill functionality.
These three skills, trace, dataset, evaluator, are not just tools; they are the foundational workflow of AI engineering, now portable. Any coding agent that adopts them can instrument, test, and improve its own behavior directly in the repository. The improvement in Claude Code’s performance is a signal, not a ceiling.
This is the beginning of a self-improving loop between agent and environment. The repo will grow. The pattern will spread.
And the line between building an agent and teaching one will continue to blur.
Related Reading
Research & Benchmarks Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks Databricks finds multi-step agents beat single-turn RAG by 21% to 38% on STaRK
Research & Benchmarks Nvidia's DLSS 4.5 beta adds 6x Multi Frame Generation for RTX 50 GPUs
From Archives Secret meeting sees 94% approve even least‑popular AI resistance stance
From Archives AI data centers move to Arctic edge, boosting Nordic rural economies
Common Questions Answered
What are the three core skills included in the LangSmith CLI?
The LangSmith CLI introduces three portable skills: trace, dataset, and evaluator. These skills enable developers to add tracing to existing code, build datasets of examples, and evaluate agent performance across different scenarios.
How can developers access the new LangSmith skills?
The skills are available in the publicly available langsmith-skills repository, allowing teams to directly import and use them without waiting for a proprietary rollout. Any coding agent that supports skill functionality can leverage these portable capabilities.
What specific benefits does the trace skill provide for coding agents?
The trace skill allows developers to add tracing to existing code and query traces, providing clearer insight into agent execution. This capability helps developers understand and debug the internal workings of their coding agents more effectively.
Further Reading
- Introducing LangSmith Fetch: Debug agents from your terminal — LangChain Blog
- Debugging Deep Agents with LangSmith — LangChain Blog
- Top 5 CLI coding agents in 2026 — Pinggy
