OpenRouter Claude 3.7 Standard throws "Response interrupted by a tool use result." in output

[cfdude]

Which version of the app are you using?

3.7.4

Which API Provider are you using?

OpenRouter

Which Model are you using?

Claude 3.7 Sonnet

What happened?

User in Discord reports: https://discord.com/channels/1332146336664915968/1343921958668079255

After roo attempts completion, subsequent follow-up messages always show this:
Response interrupted by a tool use result. Only one tool may be used at a time and should be placed at the end of the message.

There is no effect on the usability of Roo as a result. This is not a major issue, but perhaps a sign of inefficiency somewhere?

Steps to reproduce

Relevant API REQUEST output

Additional context

The user is using Memory Bank but that has no effect on this problem. Other users report getting this same message and are not using Memory Bank.

Additional context from the user:

the problem is regarding OpenRouter Sonnet 3.7.
if i use openrouter sonnet 3.5, it doesn't occur.

for what it's worth, i've downgraded to sonnet 3.5.
my experience with claude 3.7 so far has been that it's worse than claude 3.5 for existing repos. it's:

• too chatty, causing it to duplicate lines of code. claude 3.5 had the opposite problem, where it just said // rest of the code remains the same but you could prompt away from this
• commits to its mistakes. claude 3.5 was very non-committal, always agreeing with whatever you said. claude 3.7 commits to wrong things. this compounds and ultimately delivers wrong outcomes.

[ExactDoug]

maybe a bug / server error on anthropic's side?

Happening to me even in Claude Desktop sometimes after tool use. See screenshot.

[hannesrudolph]

Explanation of "Response interrupted by a tool use result" Message

What This Message Means

The "Response interrupted by a tool use result" message appears when Roo (the AI assistant) attempts to use more than one tool in a single response or tries to continue generating text after using a tool. This is not an error, but rather an intentional safeguard in the system.

Why This Happens

This interruption enforces a critical design principle in Roo's workflow: the step-by-step execution model. Specifically:

1. One tool per message: Roo is only allowed to use one tool at a time in each message
2. Wait for results: After using a tool, Roo must wait for the tool's result before proceeding
3. No assumptions: Roo should never assume what a tool's result will be
4. Informed execution: Each action Roo takes must be informed by the results of previous actions

When Roo tries to use a second tool or continue generating content after using a tool in the same message, the system intentionally interrupts the response to maintain this structured workflow.

The Benefits of This Design

This strict sequential approach has several important advantages:

• Reliability: Ensures that each action is based on accurate information from previous steps
• Transparency: Makes it clear to you what action Roo is taking at each step
• Control: Gives you the opportunity to approve or reject each step
• Recoverability: Prevents cascading errors if an earlier step fails
• Efficiency: Reduces wasted computation on actions that might be invalid

What You Should Do

When you see this message, no action is required on your part. Roo will automatically continue with the task by:

1. Processing the result of the first tool
2. Analyzing what to do next based on that result
3. Using another tool in the next message (if needed)

This creates a clean, step-by-step conversation trail where each tool use and its result are clearly paired, making the problem-solving process more transparent and reliable.

This needs to be added to https://docs.roocode.com/