anomalyco / anomalyco/opencode
[Bug] Muse Spark 1.3 Free (Xhigh): reasoning encrypted_content error on follow-up prompts with images
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 209k
- Forks
- 27.5k
- PR merge metrics
- PR metrics pending
Description
Description
I am receiving the error: Upstream request failed: [invalid_request_error] reasoning 'encrypted_content' was not issued to this caller. This happens after 2-3 messages in a new session, maybe when attaching images to a follow-up prompt. The session has heavy tool usage (over 106,000 cache tokens read), and the context limits seem to break the token validation upstream when processing the subsequent message.
Note: Similar reports have been marked as 'closed as not planned' by automated triage, but this issue remains strictly reproducible in v1.18.30.
Plugins
None
OpenCode version
v1.18.30
Steps to reproduce
- Open a project in OpenCode Desktop (v1.18.30) using OpenCode Zen provider.
- Select the "Muse Spark 1.3 Free" model and set the variant to Xhigh.
- Send an initial prompt that triggers multiple tool calls (e.g., shell commands) and builds a large cache (100k+ tokens).
- Send a 2nd or 3rd follow-up prompt in the same session, attaching screenshots to the message.
- The upstream
encrypted_contenterror immediately occurs, blocking the response.
Screenshot and/or share link
Attached exported session logs:
new-session---2026-09-14t05-56-06-838z.json
Operating System
Windows 11
Terminal
OpenCode Built-in Terminal (PowerShell)
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 failure in OpenCode Desktop v1.18.30 with the OpenCode Zen provider, Muse Spark 1.3 Free at Xhigh, heavy tool usage, and an image on a follow-up prompt. Inspect the exported session logs and the request path handling reasoning encrypted_content and context limits. Done means the second or third follow-up completes without the upstream encrypted_content error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- ai, api, desktop
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100