Mateo/dev 25 write connecting arcade tools to your llm page #595

torresmateo · 2025-12-10T21:26:41Z

Preview here: https://docs-git-mateo-dev-25-write-connecting-arcade-a84eaa-arcade-ai.vercel.app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python

Note

Introduces a concise Python guide for integrating Arcade tools with an LLM.

New guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx with step-by-step setup: project init, env vars, retrieving formatted tool definitions, auth+execute helper, multi-turn ReAct-style loop (invoke_llm), and interactive chat() example using OpenRouter
Adds the page to agent-frameworks navigation via _meta.tsx
Updates public/llms.txt to index/link the new "Connect Arcade to your LLM" doc

^{Written by Cursor Bugbot for commit 5f501dd. This will update automatically on new commits. Configure here.}

vercel · 2025-12-10T21:26:45Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Review	Updated (UTC)
docs	Ready	Preview, Comment	Jan 19, 2026 8:38pm

cursor · 2025-12-10T21:28:55Z

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

+                    "content": tool_result,
+                })
+
+            continue


Bug: Missing assistant message before tool results in history

When the LLM returns tool calls, the code appends tool result messages to history but never appends the assistant message that contained the tool_calls. The OpenAI API (and compatible APIs like OpenRouter) requires the assistant message with tool_calls to appear in the conversation history before the corresponding tool result messages. This will cause an API error on the next iteration of the loop when the malformed history is sent back to the model.

Additional Locations (1)

app/en/home/connect-arcade-to-your-llm/page.mdx#L386-L400

nearestnabors

I need more openrouter tokens to actually test that it works, but wanted to share what I have so far!

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

app/en/home/connect-arcade-to-your-llm/page.mdx

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

app/en/home/connect-arcade-to-your-llm/page.mdx

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

Co-authored-by: RL "Nearest" Nabors <[email protected]>

cursor · 2025-12-16T18:03:21Z

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

+
+        # Print the latest assistant response
+        assistant_response = history[-1]["content"]
+        print(f"\n🤖 Assistant: {assistant_response}\n")


Bug: Tool result shown as assistant response when max_turns exhausted

When invoke_llm exhausts max_turns while the assistant is still making tool calls, the function returns with a tool response as the last history item. The chat() function then accesses history[-1]["content"] and prints it prefixed with "🤖 Assistant:", displaying raw tool output as if it were the assistant's response. This produces confusing output when many consecutive tool calls are needed.

Additional Locations (1)

app/en/home/connect-arcade-to-your-llm/page.mdx#L443-L446

nearestnabors

Still getting the same malformed directory structure after running the commands in step 1:

app/en/home/connect-arcade-to-your-llm/page.mdx

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

app/en/home/connect-arcade-to-your-llm/page.mdx

Co-authored-by: vfanelle <[email protected]>

torresmateo · 2025-12-19T17:13:03Z

holding the merge until nav is merged
FYI @nearestnabors

…o-your-llm-page

cursor · 2026-01-09T02:04:25Z

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

+
+        # Print the latest assistant response
+        assistant_response = history[-1]["content"]
+        print(f"\n🤖 Assistant: {assistant_response}\n")


Max turns exceeded causes tool result printed as response

Medium Severity

The chat() function assumes history[-1] is always an assistant message and prints its content as the assistant's response. However, when invoke_llm exits because max_turns is exceeded while still processing tool calls, history[-1] is a tool message with role: "tool". This causes the raw tool result (likely a JSON string) to be displayed to the user as "🤖 Assistant:" output, creating confusing behavior.

🔬 Verification Test

Why verification test was not possible: This edge case requires the LLM to continuously make tool calls until the max_turns limit (5) is reached without providing a final response. This is difficult to trigger reliably in testing as it depends on LLM behavior and requires actual API credentials. The logic flaw is evident from code inspection: when the while loop exits due to turns >= max_turns during tool processing, no assistant message is appended, yet chat() unconditionally treats history[-1]["content"] as the assistant response.

Additional Locations (1)

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx#L460-L463

cursor

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

cursor · 2026-01-19T20:30:22Z

app/en/guides/agent-frameworks/langchain/use-arcade-with-langchain/page.mdx

+        // Handle all interrupts
+        const decisions: any[] = [];
+        for (const interrupt of interrupts) {
+          decisions.push(await handleAuthInterrupt(interrupt, rl));


Function called with extra parameter it doesn't accept

Medium Severity

The handleAuthInterrupt function is defined with a single parameter (interrupt: Interrupt) at lines 297-321, but it's called with two parameters handleAuthInterrupt(interrupt, rl) at lines 414 and 718. The rl parameter (readline interface) is passed but never used, suggesting missing functionality in the function definition.

Additional Locations (1)

app/en/guides/agent-frameworks/langchain/use-arcade-with-langchain/page.mdx#L717-L718

cursor · 2026-01-19T20:30:22Z

app/_components/early-access-registry-survey.tsx

  useEffect(() => {
-    if (!posthog) {
-      return;
-    }


PostHog initialization removed but components still use it

High Severity

The app/_components/posthog.tsx file containing posthog.init() and PostHogProvider is completely deleted, but multiple components still import and use posthog directly from posthog-js. Without initialization, posthog.capture() calls will silently fail (no analytics), and posthog.onFeatureFlags() and posthog.getSurveys() in EarlyAccessRegistrySurvey will not work, breaking the survey functionality entirely.

…o-your-llm-page

github-actions

Style Review

Found 12 style suggestion(s).

Powered by Vale + Claude

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

github-actions · 2026-01-19T20:31:34Z

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

+
+### Write a helper function that handles the LLM's invocation
+
+There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.


write-good.ThereIs: Removed 'There are' and changed 'we will' to 'you will'

Suggested change

There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.

Many orchestration patterns handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern you will implement in this example.

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

github-actions

Style Review

Found 1 style suggestion(s).

Powered by Vale + Claude

github-actions · 2026-01-19T20:34:43Z

app/en/guides/agent-frameworks/setup-arcade-with-your-llm-python/page.mdx

+
+### Write a helper function that handles the LLM's invocation
+
+There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.


write-good.ThereIs: Removed 'There are' sentence starter and changed 'we will implement' to 'you will implement' per style guide

Suggested change

There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.

Many orchestration patterns can handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern you will implement in this example.

torresmateo added 4 commits December 6, 2025 17:12

initial stab

7f25c13

prereqs

c7a1a31

added OpenRouter tutorial for wiring Arcade with any LLM

e61487f

added example code at the bottom of the guide

c645390

torresmateo requested review from nearestnabors and vfanelle December 10, 2025 21:26

cursor bot reviewed Dec 10, 2025

View reviewed changes

vercel bot deployed to Preview December 10, 2025 21:33 View deployment

nearestnabors requested changes Dec 16, 2025

View reviewed changes

torresmateo and others added 2 commits December 16, 2025 14:58

Apply suggestions from code review

57ef311

Co-authored-by: RL "Nearest" Nabors <[email protected]>

Apply suggestions from code review

1065b5d

Co-authored-by: RL "Nearest" Nabors <[email protected]>

cursor bot reviewed Dec 16, 2025

View reviewed changes

vercel bot deployed to Preview December 16, 2025 18:06 View deployment

addressed nabors' feedback

af24f12

vercel bot deployed to Preview December 16, 2025 20:12 View deployment

torresmateo requested a review from nearestnabors December 16, 2025 20:13

nearestnabors requested changes Dec 17, 2025

View reviewed changes

vfanelle requested changes Dec 18, 2025

View reviewed changes

client to arcade_client

2d401ce

vercel bot deployed to Preview December 18, 2025 13:14 View deployment

Apply suggestions from code review

f23fa7f

Co-authored-by: vfanelle <[email protected]>

vercel bot deployed to Preview December 18, 2025 13:22 View deployment

implemented val's suggestions

c707a66

torresmateo requested review from nearestnabors and vfanelle December 18, 2025 13:26

vercel bot deployed to Preview December 18, 2025 13:28 View deployment

nearestnabors approved these changes Dec 18, 2025

View reviewed changes

torresmateo enabled auto-merge (squash) December 19, 2025 16:44

torresmateo disabled auto-merge December 19, 2025 17:10

vfanelle approved these changes Dec 19, 2025

View reviewed changes

torresmateo and others added 4 commits January 8, 2026 22:49

merge main into this branch

66263c4

moved the guide to the frameworks section

5fd29f2

Merge branch 'main' into mateo/dev-25-write-connecting-arcade-tools-t…

3da6520

…o-your-llm-page

🤖 Regenerate LLMs.txt

120198a

vercel bot deployed to Preview January 9, 2026 02:04 View deployment

cursor bot reviewed Jan 9, 2026

View reviewed changes

merge with main

58a18a7

cursor bot reviewed Jan 19, 2026

View reviewed changes

Merge branch 'main' into mateo/dev-25-write-connecting-arcade-tools-t…

a65719c

…o-your-llm-page

github-actions bot reviewed Jan 19, 2026

View reviewed changes

Apply suggestions from code review

5f501dd

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

github-actions bot reviewed Jan 19, 2026

View reviewed changes

vercel bot deployed to Preview January 19, 2026 20:38 View deployment

torresmateo merged commit 585fdb3 into main Jan 19, 2026
7 checks passed

torresmateo deleted the mateo/dev-25-write-connecting-arcade-tools-to-your-llm-page branch January 19, 2026 20:46

github-actions bot mentioned this pull request Jan 19, 2026

Editorial improvements for #595 #671

Open


		### Write a helper function that handles the LLM's invocation

		There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.

	There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.
	Many orchestration patterns handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern you will implement in this example.

	There are many orchestration patterns that can be used to handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern we will implement in this example.
	Many orchestration patterns can handle the LLM invocation. A common pattern is a ReAct architecture, where the user prompt will result in a loop of messages between the LLM and the tools, until the LLM provides a final response (no tool calls). This is the pattern you will implement in this example.

Mateo/dev 25 write connecting arcade tools to your llm page #595

Mateo/dev 25 write connecting arcade tools to your llm page #595

Uh oh!

Conversation

torresmateo commented Dec 10, 2025 • edited by cursor bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vercel bot commented Dec 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cursor bot Dec 10, 2025

Choose a reason for hiding this comment

Bug: Missing assistant message before tool results in history

Uh oh!

nearestnabors left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

cursor bot Dec 16, 2025

Choose a reason for hiding this comment

Bug: Tool result shown as assistant response when max_turns exhausted

Uh oh!

nearestnabors left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

torresmateo commented Dec 19, 2025

Uh oh!

cursor bot Jan 9, 2026

Choose a reason for hiding this comment

Max turns exceeded causes tool result printed as response

Uh oh!

cursor bot left a comment

Choose a reason for hiding this comment

Uh oh!

cursor bot Jan 19, 2026

Choose a reason for hiding this comment

Function called with extra parameter it doesn't accept

Uh oh!

cursor bot Jan 19, 2026

Choose a reason for hiding this comment

PostHog initialization removed but components still use it

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Style Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot left a comment

torresmateo commented Dec 10, 2025 •

edited by cursor bot

Loading

vercel bot commented Dec 10, 2025 •

edited

Loading