🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
fix(assistant): prefer MCP tools for data analysis (#16344)
Ben Bachem112bf971bb62
diff --git a/packages/shared/src/in-app-agent/server/systemPrompt.ts b/packages/shared/src/in-app-agent/server/systemPrompt.tsindex f7a1d5d584..470b932227 100644--- a/packages/shared/src/in-app-agent/server/systemPrompt.ts+++ b/packages/shared/src/in-app-agent/server/systemPrompt.ts@@ -28,6 +28,7 @@ IMPORTANT: You should NOT answer with unnecessary preamble or postamble (such as</behavioral_rules><tools>+Prefer Langfuse MCP tools over sandbox tools whenever an MCP tool can perform the task, especially the metrics tools for aggregations.Use the docs tools to find relevant general information about Langfuse or best practices.For questions about best practices or specific workflows (e.g. setting up evals), use the available Langfuse skills when they apply.If the skill ever tells you to reference the docs, always use the langfuse-docs-mcp server to access those docs.