anomalyco / anomalyco/opencode
Desktop chat outputs corrupted internal parameter/tool text and gets stuck in Thinking
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 209k
- Forks
- 27.5k
- PR merge metrics
- PR metrics pending
Description
Description
OpenCode Desktop is showing severely corrupted assistant output during a chat.
Instead of normal assistant text, the chat starts displaying repeated internal-looking strings such as:
parameterparameterparameter...
<commentary.bat>
along with malformed XML/parameter-like content and unexpected Chinese characters.
The agent then remains in the Thinking state and does not properly complete the task. The task panel remains at 0 of 7 todos completed.
This happened while OpenCode was inspecting files and working on a website task. The malformed output appears to be generated/rendered directly into the chat UI rather than being hidden internal tool/agent data.
Expected behavior:
Internal tool/agent protocol data should not be displayed as raw corrupted text. The assistant should continue normally or show a proper error if the model response cannot be parsed.
Actual behavior:
The chat fills with repeated parameter... text and malformed tags, then appears stuck in Thinking.
OS: Windows 11
OpenCode: 1.18.25
Model: Big Pickle
Desktop app: Yes
Plugins
idk i am using opencode gui version
OpenCode version
1.18.25
Steps to reproduce
idk
Screenshot and/or share link
Operating System
windows 11
Terminal
windows
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the corrupted chat output in the OpenCode Desktop app on Windows 11 with version 1.18.25, using the reported website-inspection task if possible. Inspect the chat rendering and assistant-response handling involved when the agent remains in Thinking; done means internal parameter or tool data is hidden, malformed responses produce a proper error, and the task no longer remains stuck.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- desktop, frontend
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100