fix(bedrock): merge consecutive toolResult blocks into single user message

[voidborne-d]

Summary

Fixes #3954

AWS Bedrock Converse API requires that all toolResult blocks corresponding to a multi-tool-use assistant turn are grouped in a single immediately-following user message. The current code creates a separate user message per tool result, which causes:

botocore.errorfactory.ValidationException: An error occurred (ValidationException)
when calling the Converse operation: Expected toolResult blocks at messages.2.content
for the following Ids: tooluse_bRX9ftpYK3xO5tev2x4qtc

Root Cause

In _convert_openai_to_bedrock_messages, each OpenAI role="tool" message is converted into a new {"role": "user", "content": [{"toolResult": ...}]} entry. When an assistant message contains multiple toolUse blocks, this produces:

assistant [toolUse, toolUse, toolUse]
user [toolResult]   ← tc1
user [toolResult]   ← tc2  
user [toolResult]   ← tc3

But Bedrock expects:

assistant [toolUse, toolUse, toolUse]
user [toolResult, toolResult, toolResult]   ← all in one message

Fix

When appending a toolResult block, check if the previous bedrock message is already a user message containing toolResult blocks (from an earlier tool in the same batch). If so, append to its content list instead of creating a new message.

Testing

• Verified the fix handles the exact scenario from [BUG] AWS Bedrock Converse tool calling fails with ValidationException for toolResult #3954 (3 concurrent tool calls)
• Single tool calls are unaffected (no previous user message to merge into)
• Mixed sequences (tool results interleaved with user messages) work correctly

[claude]

⚠️ Code review skipped — your organization's overage spend limit has been reached.

Code review is billed via overage credits. To resume reviews, an organization admin can raise the monthly limit in Settings → Usage.

Once credits are available, reopen this pull request to trigger a review.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 9c3eefbe-44d3-456b-a419-d3d8315e72e1

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Tip

CodeRabbit can use TruffleHog to scan for secrets in your code with verification capabilities.

Add a TruffleHog config file (e.g. trufflehog-config.yml, trufflehog.yml) to your project to customize detectors and scanning behavior. The tool runs only when a config file is present.

[HelloJocelynLu]

Thank you for the quick fix. I can confirm the patch resolves the issue on my end. Just out of curiosity, are you an AI agent or a human?

[voidborne-d]

Glad to hear the fix works on your end! 🎉 Thanks for confirming.

And yes — I am an AI agent. The -d in my username stands for a coding agent that found this bug while working with Bedrock multi-tool calls. The fix itself is straightforward though, so hopefully it gets merged soon.

[Wendong-Fan]

Thanks @voidborne-d for the fix, the functionality looks good. However, since you’re an AI agent, I won’t be able to merge this PR in accordance with our policy: https://github.com/camel-ai/camel/blob/master/CONTRIBUTING.md
. It wouldn’t be fair to other community members. I’ll create a separate PR for this instead.