fix(session): recover stale encrypted reasoning by 1837620622 · Pull Request #48918 · anomalyco/opencode

1837620622 · 2026-09-14T03:50:35Z

Issue for this PR

Type of change

Bug fix
New feature
Refactor / code improvement
Documentation

What does this PR do?

Resuming a Muse Spark / OpenAI Responses session after idle fails with:

reasoning `encrypted_content` was not issued to this caller

OpenCode stored that caller-bound blob and treated HTTP 400 as fatal, so the session stayed bricked.

This follows the existing context-overflow recovery path:

Classify those errors as stale-reasoning
If no assistant output has started, drop reasoningEncryptedContent / itemId from stored reasoning
Retry the turn once

Readable reasoning text stays. Live tool-loop encrypted state is left alone unless the provider rejects it.

Related: #48741, #48805

How did you verify your code works?

packages/llm: bun test test/provider-error.test.ts
packages/opencode: bun test test/provider/error.test.ts and the fromError case in test/session/message-v2.test.ts
packages/core: bun test test/session-runner.test.ts (90 pass), including three new recovery tests plus existing overflow and encrypted-reasoning tests

Screenshots / recordings

n/a (non-UI)

Checklist

I have tested my changes locally
I have not included unrelated changes in this PR

Retry once after stripping caller-bound reasoning.encrypted_content when the provider rejects it, matching overflow recovery.

github-actions · 2026-09-14T03:51:13Z

The following comment was made by an LLM, it may be inaccurate:

Potential Duplicate Found:

PR #48908 - fix(session): recover from stale encrypted reasoning on provider rejection

fix(session): recover from stale encrypted reasoning on provider rejection #48908

Why it might be related:
This PR addresses the same issue with stale encrypted reasoning recovery. Both PRs appear to handle the recovery path when encrypted reasoning content becomes stale during session resumption. They likely overlap in implementation or address the same bug from different angles.

Note: PR #43595 fix(opencode): recover stale encrypted reasoning state is an older related PR on this topic, but may be superseded.

1837620622 · 2026-09-14T03:52:04Z

This bricks Muse Spark 1.3 on 1.18.30 when resuming an old session. Tests in packages/core session-runner pass. Please review/merge when you can.

SynacNipo · 2026-09-14T03:58:14Z

hopefully

…ning

kungfooman · 2026-09-15T13:25:38Z

I got that error message too now, what helps is using /models to e.g. switch from 1.2 to 1.3 or vice versa. So it's not a "full" brick.

yakovexplorer · 2026-09-15T14:57:31Z

still persists on 1.18.31. I hope this gets merged soon.

kvnloo · 2026-09-15T18:16:57Z

Why this matters

Resuming after idle still bricks sessions with stale encrypted_content / invalid_encrypted_content on 1.18.31 (see thread + #48915). Two open PRs both try to recover — maintainers need one shape on dev, not a third.

Evidence

Two open recoveries for stale encrypted reasoning (encrypted_content not issued to this caller / invalid_encrypted_content):

	#48918	#48908
Base	`dev`	`v2` (currently conflicting)
Tip	`87dd316`	`73e8e76`
Error class	`packages/llm` + opencode provider error	`packages/ai`
Durability	strip + retry (also touches message-v2 / processor)	strip + `SessionEvent.MessageContentUpdated` so projection replay keeps sanitized state

Community still sees the brick on 1.18.31 (thread on #48918).

Architecture (proposal)

Maintainer call: which shape should land on dev? Happy to help rebase/tests on the chosen one — not opening a third PR.

Retry once after stripping caller-bound reasoning.encrypted_content when the provider rejects it, matching overflow recovery. Adapted from PR anomalyco#48918 by 传康Kk <1837620622@qq.com>. Fork scope: core runner + @opencode-ai/llm classifier + tests only. The packages/opencode legacy processor path from the PR is intentionally excluded (heavily diverged in this fork and out of Wave Q scope). Type-only adaptation: SessionStaleReasoning.stripOpenaiReplay is typed against the schema ProviderMetadata shape to satisfy this fork's tsgo.

Retry once after stripping caller-bound reasoning.encrypted_content when the provider rejects it, matching overflow recovery. This ports the packages/opencode side of PR anomalyco#48918 (server handler -> SessionPrompt -> SessionProcessor -> llm) that Wave Q intentionally left out. Adapted from PR anomalyco#48918 by 传康Kk <1837620622@qq.com> (V1-path port, Wave R). Fork scope: provider/error.ts classifier, session/stale-reasoning.ts sanitizer, session/processor.ts retry-once hook, message-v2.ts serialization, and tests. Adaptation: processor.ts diverged ~+329 lines from the merge-base, so the hook was mirrored onto the fork's retry/backoff + finalizeRetryInflight structure instead of cherry-picking. outputStarted()/recoverStaleReasoning follow upstream.

AceRothstein71 · 2026-09-19T15:00:29Z

+1 for shipping this — it fixes a real production brick (Muse Spark sessions failing with reasoning 'encrypted_content' was not issued to this caller after idle-resume; the 400 was fatal and the session stayed stuck).

We vendored this recovery into our fork and verified it end-to-end. One detail worth noting: live sessions still drain through the V1 path (packages/opencode → SessionProcessor → ./llm), not only the core runner — so we adopted both halves: your packages/llm classifier + core runner changes, and the packages/opencode side (provider/error.ts + session/stale-reasoning.ts + processor hook). Verified behaviors:

rejection before output → exactly one retry; replayed request stripped of reasoningEncryptedContent/itemId on reasoning parts; visible text preserved
second failure → surfaced, exactly 2 requests, no loop
rejection after assistant output started → no recovery, no strip
classifier matches the live APICallError text (... was not issued to this caller) and invalid_encrypted_content

Regression on our branch: opencode test/session/ 452/0, provider dir 722/0, core 1138/0, llm 300/0; typechecks clean. Carried in AceRothstein71/opencode session-diff-events (3d6de47fc6 core side, b25cd68045 V1 port) — happy to drop or keep it rebased once this or #48908 lands. Thanks for the fix.

fix(session): recover stale encrypted reasoning

87dd316

Retry once after stripping caller-bound reasoning.encrypted_content when the provider rejects it, matching overflow recovery.

1837620622 mentioned this pull request Sep 14, 2026

Muse Spark 1.3: resuming a session after idle fails with reasoning encrypted_content not issued to this caller #48915

Open

This was referenced Sep 14, 2026

Opencode Zen critical errors on Muse Spark family when model recieves a image/does a tool call #48741

Open

[Bug] muse-spark-1.3-contributor-free: reasoning encrypted_content was not issued to this caller when switching models in same session #48805

Open

github-actions Bot mentioned this pull request Sep 14, 2026

📊 AI CLI 工具社区动态日报 2026-09-14 96loveslife/big_model_radar#486

Open

witekjedrzejewski mentioned this pull request Sep 14, 2026

compaction: reasoning encrypted_content from the compaction agent is replayed into the main chain → permanently bricked session on Azure Foundry Responses #48942

Closed

VishvakR mentioned this pull request Sep 14, 2026

[Bug] Muse Spark 1.3 Free (Xhigh): reasoning encrypted_content error on follow-up prompts with images #48947

Open

Vargas876 mentioned this pull request Sep 15, 2026

[BUG] Console muse-spark-1.3-contributor-free: 400 'reasoning encrypted_content was not issued to this caller' (zen/v1/responses) #49116

Closed

SynacNipo added a commit to SynacNipo/opencode that referenced this pull request Sep 15, 2026

Merge PR anomalyco#48918: fix(session): recover stale encrypted reaso…

732eea0

…ning

morpheus9393 mentioned this pull request Sep 15, 2026

Zen muse-spark-*: tool-call continuation 400s with reasoning encrypted_content was not issued to this caller #49206

Closed

Navigation Menu

fix(session): recover stale encrypted reasoning - #48918

fix(session): recover stale encrypted reasoning#48918
1837620622 wants to merge 1 commit into
anomalyco:devfrom
1837620622:fix/stale-encrypted-reasoning

1837620622 commented Sep 14, 2026

Uh oh!

github-actions Bot commented Sep 14, 2026

Uh oh!

1837620622 commented Sep 14, 2026

Uh oh!

SynacNipo commented Sep 14, 2026

Uh oh!

kungfooman commented Sep 15, 2026

Uh oh!

yakovexplorer commented Sep 15, 2026

Uh oh!

kvnloo commented Sep 15, 2026

Uh oh!

AceRothstein71 commented Sep 19, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Conversation

1837620622 commented Sep 14, 2026

Issue for this PR

Type of change

What does this PR do?

How did you verify your code works?

Screenshots / recordings

Checklist

Uh oh!

github-actions Bot commented Sep 14, 2026

Uh oh!

1837620622 commented Sep 14, 2026

Uh oh!

SynacNipo commented Sep 14, 2026

Uh oh!

kungfooman commented Sep 15, 2026

Uh oh!

yakovexplorer commented Sep 15, 2026

Uh oh!

kvnloo commented Sep 15, 2026

Why this matters

Evidence

Architecture (proposal)

Uh oh!

AceRothstein71 commented Sep 19, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants