AE-H96FWV84
Browser agent stores its past tool calls as plain JSON text in history, so local models imitate that format and emit malformed tool calls
TOOL FAILUREseverity: HIGHcause: LIKELYoutcome: RESOLVED UNVERIFIEDconfidence: MEDIUM
browser-use's message manager added previous agent outputs to history as ordinary AI messages containing a JSON dump instead of structured tool calls. With Ollama models such as qwen2.5, the model copied that format rather than the required tool-call format, so its tool calls could not be parsed and the parsed output was None.
- Framework / agent
- browser-use · browser automation agent
- Remediation attempts
- APPLIEDSUGGESTED
- Recurrence
- not documented
- Source languages
- en
- Updated
- 2026-09-30
Sources
- GITHUB ISSUE [Bug] Tool Calls Incorrectly Stored as Plain Messages in Message History when using Ollama — github.com/browser-use/browser-use, retrieved 2026-09-29
- GITHUB PULL REQUEST Fix tool call message formatting and add example tool call — github.com/browser-use/browser-use, retrieved 2026-09-29
Symptoms
- The model mimics the JSON-in-text history format instead of producing proper tool calls
- Structured output parsing returns None while the raw output contains a <tool_call> block
- After the first action the model forgets the output structure
The full record — root-cause evidence, every remediation attempt with its status and verification, failed attempts, patch references, verbatim quotes and recurrence — is a paid lookup (0.018 USDC via x402). Agents:
GET /api/v1/cases/AE-H96FWV84. PricingSimilarity to your system is not implied. A remediation that worked in the documented context may not work in yours.