OpenAgentSkill Registry Manifest
Skill: langgraph-testing-evaluation
Slug: soba-labs-langgraph-testing-evaluation
Category: design-creative
Description: Use this skill when you need to test or evaluate LangGraph/LangChain agents: writing unit or integration tests, generating test scaffolds, mocking LLM/tool behavior, running trajectory evaluation (match or LLM-as-judge), running LangSmith dataset evaluations, and comparing two agent versions with A/B-style offline analysis. Use it for Python and JavaScript/TypeScript workflows, evaluator design, experiment setup, regression gates, and debugging flaky/incorrect evaluation results.

Agent fit:
- Decision: 66/100 Prototype first
- Primary fit: Research agents
- Role: Fallback candidate

Supply profile:
- Track: Coding and developer agents
- Scenario: Testing and QA
- Applicable agents: Claude Code, OpenAI Agents, LangChain, CLI, Codex
- Maintenance: 22d since push
- Risk: Needs review

Trust:
- Trust score: 77/100 Strong shortlist
- Audit: 80/100 Needs review

Attribution:
- Status: Registry indexed
- Source: github candidate review
- Creator: soba-labs
- Claim URL: https://www.openagentskill.com/skills/soba-labs-langgraph-testing-evaluation#claim-this-skill

Install:
npx skills add soba-labs/langchain-agent-skills --skill langgraph-testing-evaluation

URLs:
- Web: https://www.openagentskill.com/skills/soba-labs-langgraph-testing-evaluation
- API: https://www.openagentskill.com/api/agent/skills/soba-labs-langgraph-testing-evaluation
- Install API: https://www.openagentskill.com/api/skills/soba-labs-langgraph-testing-evaluation/install
- Repository: https://github.com/soba-labs/langchain-agent-skills/tree/main/skills/langgraph-testing-evaluation