Show what a LangChain model call was sent, in LangChain's words

[wu-sheng]

The page's Prompt tab read a LangChain request as Claude Code's shape and drew it as one empty message. The fix is in Horizon's renderer (apache/skywalking-horizon-ui#172). This moves the embedded renderer to that commit.

• internal/view/conversation-view is rebuilt with tools/conversation-view.sh update 297f0205e502a3bc795d236356a31364dec65b5c. That is the newest commit on Horizon's main.
• Between the old pin (53c1f63) and the new one, two Horizon commits touch what asz embeds:
  • #172 changes how the Prompt tab reads LangChain bodies.
  • #167 changes two rules in the theme: a light theme (Daybreak) draws scrollbars and other native controls light, and the top bar's actions give way on a narrow window. asz's pages set data-appearance, so the first rule applies here.
• docs/en/changes/changes.md records the change under LangChain and LangGraph.

What the page shows now for a LangChain call

• Each message has LangChain's own role: system, human, ai or tool.
• An ai message shows its text, then its tool calls, each drawn once.
• A tool message shows its result under the call it answers.
• A response shows the answer, the model and the token counts.
• What it added stays unavailable for a LangChain call, because no request names the one before it.

Checked

• make conversation-view-check passes: the embedded copy is what 297f020 builds.
• make check passes.
• I built asz from this branch, served a real LangGraph conversation (SummarizationMiddleware on Amazon Bedrock) and opened a call's Prompt tab. It shows the system prompt, the middleware's summary as a human message, then the ai message with its tool call.