PR #59643: fix(agents): preserve commentary/final_answer phase separation

• Repository: openclaw/openclaw
• Author: ringlochid
• State: closed | merged: True
• Link: https://github.com/openclaw/openclaw/pull/59643

Description (problem / solution / changelog)

Summary

• Problem: assistant turns that contain both commentary and final_answer text can be flattened into one visible output, which leaks commentary into user-facing replies and can produce duplicate or malformed final delivery.
• Why it matters: this breaks the expected final-only user experience, causes duplicate replies after tool/send paths, and corrupts replay/context because mixed-phase text is persisted and replayed ambiguously.
• What changed: preserved phase separation end-to-end across stored-message conversion, replay/input-item rebuilding, WebSocket partial phase propagation, and visible extraction/delivery so user-visible output prefers final_answer while still falling back safely when no final text exists.
• What did NOT change (scope boundary): this does not globally redefine every text extractor in the repo, does not change tool-call semantics, and does not attempt a broader phase-aware audit outside the main OpenAI WS -> embedded subscribe -> visible delivery path.

Change Type (select all)

• Bug fix
• Feature
• Refactor required for the fix
• Docs
• Security hardening
• Chore/infra

Scope (select all touched areas)

• Gateway / orchestration
• Skills / tool execution
• Auth / tokens
• Memory / storage
• Integrations
• API / contracts
• UI / DX
• CI/CD / infra

Linked Issue/PR

• Closes #59150
• Related #56198
• Related #58892
• Related #52084
• Related #25592
• Related #44467
• Related PR #30479
• Related PR #57484
• This PR fixes a bug or regression

Root Cause / Regression History (if applicable)

• Root cause: assistant text phase truth already existed at block level, but several layers still flattened mixed-phase text. Stored assistant messages could carry both commentary and final-answer blocks under one misleading top-level phase; replay collapsed them back together; stream partials could lose item-phase attribution; and visible extraction/delivery then consumed flattened text instead of phase-aware text.
• Missing detection / guardrail: there was no focused regression coverage for mixed-phase stored messages, phase-aware replay splitting, signature-only phased partials, or text_end / message_end delivery interactions where commentary previews must be suppressed/replaced by final output.
• Prior context (git blame, prior PR, issue, or refactor if known): this bug family overlaps longstanding leakage/duplication reports in #59150, #56198, #58892, #52084, #25592, and #44467. Adjacent prior art includes #30479 (stripping raw user-facing protocol leakage) and #57484 (commentary-delivery semantics on a channel path), but those did not fix mixed-phase persistence/replay/delivery end-to-end.
• Why this regressed now: the issue is not a single recent regression; it is an accumulated phase-separation gap that became more visible once commentary, block replies, tool sends, and final-answer delivery all coexisted on the same assistant turn path.
• If unknown, what was ruled out: ruled out “display-only” root cause. Investigation confirmed the problem begins upstream in stored-message conversion/replay semantics, not only in final rendering.

Regression Test Plan (if applicable)

• Coverage level that should have caught this:
  • Unit test
  • Seam / integration test
  • End-to-end test
  • Existing coverage already sufficient
• Target test or file:
  • src/agents/openai-ws-stream.test.ts
  • src/agents/pi-embedded-utils.test.ts
  • src/agents/pi-embedded-subscribe.handlers.messages.test.ts
  • src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.emits-block-replies-text-end-does-not.test.ts
  • src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.does-not-append-text-end-content-is.test.ts
  • src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.does-not-duplicate-text-end-repeats-full.test.ts
• Scenario the test should lock in: mixed commentary/final stored messages stay phase-separated on replay; stream partials preserve phase; visible extraction prefers final_answer -> commentary -> legacy/unphased; commentary text_end block replies are suppressed until final delivery; final replacement at message_end works; and duplicate/prefix-extension regressions remain fixed.
• Why this is the smallest reliable guardrail: the bug spans stored-message conversion, stream partial attribution, and delivery seams. Unit-only coverage at one layer would miss the cross-layer collapse that caused the visible duplication/leak.
• Existing test that already covers this (if any): none before this branch for the mixed-phase end-to-end path.
• If no new test is added, why not: N/A

User-visible / Behavior Changes

• Visible assistant output now prefers final_answer text when both commentary and final-answer phases exist in one turn.
• Commentary-only previews are no longer allowed to leak through as the final visible reply in the main embedded delivery path.
• When commentary streamed first and final text arrives later, the final visible reply replaces the preview instead of duplicating it.
• If no final-answer text exists, commentary/unphased fallback still works instead of producing an empty reply.

Diagram (if applicable)

Before:
[mixed commentary + final_answer blocks]
  -> [stored/replayed as flattened assistant text]
  -> [visible extractor concatenates all text]
  -> [commentary leak and/or duplicate final reply]

After:
[mixed commentary + final_answer blocks]
  -> [phase preserved in storage/replay/partials]
  -> [visible extractor prefers final_answer]
  -> [commentary preview suppressed/replaced]
  -> [single intended final visible reply]

Security Impact (required)

• New permissions/capabilities? No
• Secrets/tokens handling changed? No
• New/changed network calls? No
• Command/tool execution surface changed? No
• Data access scope changed? No
• If any Yes, explain risk + mitigation: N/A

Repro + Verification

Environment

• OS: Ubuntu (local dev host)
• Runtime/container: local Node/pnpm repo checkout
• Model/provider: OpenAI WS / embedded subscribe path
• Integration/channel (if any): embedded delivery path with Telegram/Discord-adjacent visible-output semantics
• Relevant config (redacted): default OpenAI WS / embedded subscribe test harnesses; no special secrets required

Steps

1. Produce or replay an assistant turn containing both commentary and final_answer text blocks.
2. Observe stored/replayed assistant content and the user-visible delivery path.
3. Confirm whether visible output leaks commentary, duplicates final delivery, or collapses mixed phases.

Expected

• Stored and replayed assistant content preserves phase boundaries.
• User-visible extraction prefers final_answer when present.
• Commentary previews are suppressed or replaced rather than duplicated.

Actual

• Before this fix: mixed-phase turns could flatten into one visible assistant reply, leak commentary, or produce duplicate final delivery.
• On this branch: phase separation is preserved through replay and delivery, and the targeted duplicate/leak regressions are covered by tests.

Evidence

Attach at least one:

• Failing test/log before + passing after

• Trace/log snippets

• Screenshot/recording

• Perf numbers (if relevant)

• local transcript evidence showed assistant turns with both commentary and final_answer blocks in a single message

• targeted regression slice passed: 8 suites / 164 tests

• fresh independent review loop passed for storage/replay, stream-phase, visible-delivery, and holistic closure before branch submission

Human Verification (required)

What you personally verified (not just CI), and how:

• Verified scenarios:
  • inspected mixed-phase transcript evidence and confirmed commentary + final-answer coexistence in one assistant turn
  • verified stored-message/replay, stream partial propagation, and visible delivery changes in the touched files
  • ran the targeted regression slice and confirmed 8 suites / 164 tests passed
  • confirmed the branch diff remains scoped to the intended 10 files
• Edge cases checked:
  • commentary leaking through text_end block replies
  • final replacement at message_end after commentary streamed first
  • legacy/unphased + phased replay collapse
  • signature-only phased partials without top-level partial.phase
  • prefix-extension and duplicate text_end regressions
• What you did not verify:
  • full-repo tsc --noEmit on this host (prior typecheck attempts were memory-constrained)
  • a broader follow-up audit of other phase-blind helper paths such as src/agents/tools/sessions-helpers.ts

Review Conversations

• I replied to or resolved every bot review conversation I addressed in this PR.
• I left unresolved only the conversations that still need reviewer or maintainer judgment.

Compatibility / Migration

• Backward compatible? Yes
• Config/env changes? No
• Migration needed? No
• If yes, exact upgrade steps: N/A

Risks and Mitigations

• Risk: other assistant-text consumers outside the main embedded delivery path may still use phase-blind flattening and show adjacent inconsistencies.
  • Mitigation: this PR keeps scope tight to the verified main issue path and leaves sessions-helpers parity as explicit follow-up watchpoint rather than silently broadening behavior.
• Risk: replay/delivery edge cases could regress around partial/final transitions.
  • Mitigation: regression coverage now locks in mixed-phase replay splitting, phase-aware partial attribution, commentary suppression, final replacement, and duplicate/text-end edge cases.

Changed files

• CHANGELOG.md (modified, +1/-0)
• src/agents/openai-ws-message-conversion.ts (modified, +35/-13)
• src/agents/openai-ws-stream.test.ts (modified, +490/-30)
• src/agents/openai-ws-stream.ts (modified, +145/-12)
• src/agents/pi-embedded-subscribe.handlers.messages.test.ts (modified, +169/-0)
• src/agents/pi-embedded-subscribe.handlers.messages.ts (modified, +84/-32)
• src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.does-not-append-text-end-content-is.test.ts (modified, +1/-0)
• src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.does-not-duplicate-text-end-repeats-full.test.ts (modified, +4/-1)
• src/agents/pi-embedded-subscribe.subscribe-embedded-pi-session.emits-block-replies-text-end-does-not.test.ts (modified, +468/-1)
• src/agents/pi-embedded-utils.test.ts (modified, +76/-1)
• src/agents/pi-embedded-utils.ts (modified, +108/-6)

PR #61457: fix(agents): suppress commentary partial visibility and buffer unphased early deltas

• Repository: openclaw/openclaw
• Author: 100yenadmin
• State: closed | merged: False
• Link: https://github.com/openclaw/openclaw/pull/61457

Description (problem / solution / changelog)

Summary

Addresses the two specific blockers from the maintainer review on #59643, building on the architectural direction of that PR while fixing the behaviors that blocked its merge.

Blocker 1: Commentary partials must stay suppressed

The merged fix from #61282 already suppresses commentary-phase output at the embedded subscribe delivery boundary on current main. This PR preserves that behavior and does not reintroduce the commentary-preview-then-replace pattern that was rejected in #59643.

Blocker 2: Late-map unphased deltas must be buffered

In src/agents/openai-ws-stream.ts, when response.output_text.delta arrives before response.output_item.added, the delta is now buffered instead of emitted as an unphased visible partial. When item metadata arrives:

• final_answer buffered text is emitted with correct phase/signature
• commentary buffered text stays suppressed

Architectural improvements (from #59643 direction)

• Phase-aware replay conversion in openai-ws-message-conversion.ts: stored/replayed assistant text respects per-block textSignature.phase and splits by phase instead of flattening
• Phase attribution on WS partials via textSignature
• Top-level assistant phase only set when response is consistently single-phase

What this does NOT cover (known follow-up scope)

• src/agents/tools/sessions-helpers.ts still uses phase-blind extractAssistantText() — follow-up PR recommended
• TUI render path (tui-formatters.ts, tui-stream-assembler.ts) still phase-blind — follow-up PR recommended
• Webchat/dashboard history consumers may still render generic assistant text — follow-up PR recommended

These are explicitly out of scope per the maintainer direction to make the core WS → embedded subscribe path correct first.

Change Type

• Bug fix

Scope

• Gateway / orchestration
• Integrations

Linked Issues

• Closes #59150
• Related #59643
• Related #61282

Tests

Tests added for:

• Phase-aware replay splitting
• Mixed-phase stored message reconstruction
• WS partial emission only after item-phase mapping exists
• Suppression of buffered commentary deltas when late phase resolves to commentary

Credits

Architectural direction from @ringlochid in #59643. Review feedback from @steipete that identified the two specific blockers addressed here.

Changed files

• src/agents/openai-ws-message-conversion.ts (modified, +34/-13)
• src/agents/openai-ws-stream.test.ts (modified, +543/-0)
• src/agents/openai-ws-stream.ts (modified, +173/-12)
• src/agents/pi-embedded-subscribe.handlers.messages.test.ts (modified, +180/-0)
• src/agents/pi-embedded-subscribe.handlers.messages.ts (modified, +14/-0)

PR #61463: fix(agents,gateway): phase-aware assistant text extraction — suppress OpenAI commentary leaks in sessions-helpers, TUI, and REST history

• Repository: openclaw/openclaw
• Author: 100yenadmin
• State: closed | merged: True
• Link: https://github.com/openclaw/openclaw/pull/61463

Description (problem / solution / changelog)

What this fixes (plain English)

Several places in the codebase that display assistant text were not aware of the phase system (commentary vs. final answer). This meant internal "thinking out loud" commentary could leak into user-visible surfaces like the TUI and HTTP/SSE session history endpoints. This PR makes those surfaces phase-aware, and hardens heartbeat session resolution against subagent key leaks.

Technical details

Follow-up to #59643 and #59150, which fixed phase separation in the core WS path. Post-merge audit found adjacent surfaces still using phase-blind extraction.

Surfaces fixed:

• src/tui/tui-formatters.ts — assistant text extraction now uses extractAssistantVisibleText() to prefer final_answer over commentary phase blocks
• src/infra/heartbeat-runner.ts — heartbeat session resolution now rejects subagent session keys, falling back to the main session key instead of leaking into subagent scopes
• src/gateway/sessions-history-http.test.ts — regression test: REST history applies chat.history sanitization (strips NO_REPLY messages, preserves phase blocks)
• src/tui/tui-formatters.test.ts — regression test: mixed commentary + final_answer blocks extract only the visible final answer

Explicitly deferred: extractAssistantTextForSilentCheck and buffered delta/final rendering — lower-confidence, more behaviorally sensitive.

Related

• Follow-up to #59643 and #59150
• Follow-up issues: #61474, #61475, #61476, #61477, #61478
• Companion PRs: #61529, #61528, #61527
• Related PRs: #61855, #61816

Test plan

• 27/27 TUI formatter tests pass
• Regression test for mixed commentary + final_answer in TUI extraction
• HTTP history regression test (REST path shares chat.history sanitization, strips NO_REPLY messages)
• SSE seq validation for NO_REPLY message stripping
• sessions-history-http gateway integration tests have 2 pre-existing infra failures (also fail on upstream/main)

Changed files

• CHANGELOG.md (modified, +1/-0)
• src/infra/heartbeat-runner.subagent-session-guard.test.ts (added, +72/-0)
• src/infra/heartbeat-runner.ts (modified, +24/-17)
• src/tui/tui-formatters.test.ts (modified, +20/-0)
• src/tui/tui-formatters.ts (modified, +25/-0)

PR #61481: fix(agents): harden OpenAI phase-aware visible text — suppress commentary partials, prevent empty final_answer fallback leak

• Repository: openclaw/openclaw
• Author: 100yenadmin
• State: closed | merged: False
• Link: https://github.com/openclaw/openclaw/pull/61481

Description (problem / solution / changelog)

Summary

• fix phase-aware visible text extraction so an explicit final_answer block never falls back to commentary or legacy unphased text when it sanitizes to empty
• suppress all commentary-phase partial streaming output regardless of whether extracted visible text is non-empty
• keep session-history HTTP/SSE sanitization aligned with the hardened chat history path
• add regression tests covering both leak paths and the session-history follow-through

Context

This hardens the merged #59643 behavior against two P1 leaks:

• fixes #61474
• fixes #61475

Related issues / bug family

• related to #25592
• related to #59536
• related to #59918
• related to #44213
• related to #49438
• related to #53960

Parent / sibling PRs

• parent: #59643 — core phase-separation fix (merged)
• sibling: #61463 — phase-aware extraction in sessions-helpers, TUI, and history paths

Remaining follow-ups from the same adversarial review

• #61476 — replay splitting corrupts phase on mixed messages
• #61477 — late-map buffering gates on key existence, not phase validity
• #61478 — function-call replay silently loses malformed arguments

Related open PRs

• #59920 — prefer terminal reply fields in CLI JSONL parser
• #61151 — drop partialJson streaming artifacts from session history
• #61337 — disable OpenAI tool-use pairing repair

Testing

• npm exec -- node --no-maglev ./node_modules/vitest/vitest.mjs run --config vitest.config.ts src/agents/pi-embedded-utils.test.ts src/agents/pi-embedded-subscribe.handlers.messages.test.ts
• npm exec -- node --no-maglev ./node_modules/vitest/vitest.mjs run --config vitest.config.ts src/gateway/sessions-history-http.test.ts

Changed files

• .agents/skills/openclaw-parallels-smoke/SKILL.md (modified, +13/-0)
• .agents/skills/openclaw-qa-testing/SKILL.md (added, +86/-0)
• .agents/skills/openclaw-qa-testing/agents/openai.yaml (added, +4/-0)
• .github/labeler.yml (modified, +4/-0)
• .github/workflows/ci.yml (modified, +7/-1)
• .github/workflows/control-ui-locale-refresh.yml (modified, +2/-2)
• .github/workflows/openclaw-npm-release.yml (modified, +1/-1)
• CHANGELOG.md (modified, +40/-12)
• appcast.xml (modified, +248/-116)
• apps/android/app/build.gradle.kts (modified, +2/-2)
• apps/ios/Config/Version.xcconfig (modified, +3/-3)
• apps/macos/Sources/OpenClaw/Resources/Info.plist (modified, +2/-2)
• apps/macos/Sources/OpenClawProtocol/GatewayModels.swift (modified, +14/-0)
• apps/shared/OpenClawKit/Sources/OpenClawKit/Resources/tool-display.json (modified, +23/-0)
• apps/shared/OpenClawKit/Sources/OpenClawProtocol/GatewayModels.swift (modified, +14/-0)
• docs/.generated/config-baseline.sha256 (modified, +4/-4)
• docs/.generated/plugin-sdk-api-baseline.sha256 (modified, +2/-2)
• docs/automation/tasks.md (modified, +5/-0)
• docs/channels/discord.md (modified, +1/-1)
• docs/channels/matrix.md (modified, +29/-5)
• docs/cli/memory.md (modified, +43/-15)
• docs/cli/update.md (modified, +3/-1)
• docs/concepts/dreaming.md (modified, +121/-194)
• docs/concepts/memory-qmd.md (modified, +17/-1)
• docs/concepts/memory-search.md (modified, +9/-8)
• docs/concepts/memory.md (modified, +12/-8)
• docs/concepts/model-providers.md (modified, +2/-0)
• docs/concepts/models.md (modified, +2/-0)
• docs/docs.json (modified, +8/-1)
• docs/gateway/configuration-reference.md (modified, +31/-12)
• docs/help/faq.md (modified, +36/-0)
• docs/help/testing.md (modified, +22/-0)
• docs/install/updating.md (modified, +1/-0)
• docs/plugins/architecture.md (modified, +1/-0)
• docs/plugins/building-plugins.md (modified, +1/-0)
• docs/plugins/manifest.md (modified, +76/-30)
• docs/plugins/sdk-migration.md (modified, +11/-1)
• docs/plugins/sdk-overview.md (modified, +22/-9)
• docs/providers/bedrock-mantle.md (modified, +20/-7)
• docs/providers/bedrock.md (modified, +29/-0)
• docs/providers/comfy.md (added, +201/-0)
• docs/providers/fal.md (modified, +2/-1)
• docs/providers/google.md (modified, +30/-0)
• docs/providers/index.md (modified, +4/-0)
• docs/providers/minimax.md (modified, +29/-0)
• docs/providers/models.md (modified, +4/-0)
• docs/providers/openai.md (modified, +10/-2)
• docs/providers/runway.md (added, +63/-0)
• docs/providers/vydra.md (added, +123/-0)
• docs/reference/memory-config.md (modified, +117/-98)
• docs/tools/image-generation.md (modified, +21/-17)
• docs/tools/index.md (modified, +14/-7)
• docs/tools/lobster.md (modified, +11/-9)
• docs/tools/music-generation.md (added, +208/-0)
• docs/tools/plugin.md (modified, +1/-0)
• docs/tools/slash-commands.md (modified, +1/-1)
• docs/tools/video-generation.md (modified, +147/-84)
• docs/web/control-ui.md (modified, +4/-1)
• docs/web/dashboard.md (modified, +2/-0)
• dream-diary-preview-v2.html (added, +399/-0)
• dream-diary-preview-v3.html (added, +323/-0)
• extensions/amazon-bedrock-mantle/api.ts (modified, +2/-0)
• extensions/amazon-bedrock-mantle/bedrock-token-generator.d.ts (added, +6/-0)
• extensions/amazon-bedrock-mantle/discovery.test.ts (modified, +101/-3)
• extensions/amazon-bedrock-mantle/discovery.ts (modified, +64/-13)
• extensions/amazon-bedrock-mantle/package.json (modified, +3/-0)
• extensions/bluebubbles/src/accounts.ts (modified, +5/-1)
• extensions/bluebubbles/src/monitor.ts (modified, +1/-1)
• extensions/browser/src/browser/chrome.default-browser.test.ts (modified, +2/-6)
• extensions/browser/src/browser/client-fetch.loopback-auth.test.ts (modified, +2/-6)
• extensions/browser/src/browser/control-service.plugin-disabled.test.ts (modified, +2/-6)
• extensions/browser/src/browser/profiles-service.test.ts (modified, +5/-8)
• extensions/browser/src/browser/pw-tools-core.clamps-timeoutms-scrollintoview.test.ts (modified, +2/-6)
• extensions/browser/src/browser/pw-tools-core.interactions.batch.test.ts (modified, +2/-6)
• extensions/browser/src/browser/pw-tools-core.interactions.evaluate.abort.test.ts (modified, +2/-6)
• extensions/browser/src/browser/pw-tools-core.interactions.set-input-files.test.ts (modified, +2/-4)
• extensions/browser/src/browser/pw-tools-core.last-file-chooser-arm-wins.test.ts (modified, +2/-6)
• extensions/browser/src/browser/pw-tools-core.screenshots-element-selector.test.ts (modified, +2/-6)
• extensions/browser/src/browser/routes/agent.existing-session.test.ts (modified, +3/-8)
• extensions/browser/src/browser/routes/basic.existing-session.test.ts (modified, +3/-8)
• extensions/browser/src/browser/server-context.existing-session.test.ts (modified, +3/-8)
• extensions/browser/src/browser/server-context.hot-reload-profiles.test.ts (modified, +6/-12)
• extensions/browser/src/browser/server-context.remote-profile-tab-ops.fallback.test.ts (modified, +2/-6)
• extensions/browser/src/browser/server-context.remote-profile-tab-ops.playwright.test.ts (modified, +2/-6)
• extensions/browser/src/browser/server-lifecycle.test.ts (modified, +3/-8)
• extensions/browser/src/browser/server.control-server.test-harness.ts (modified, +2/-1)
• extensions/browser/src/browser/server.evaluate-disabled-does-not-block-storage.test.ts (modified, +3/-8)
• extensions/browser/src/cli/browser-cli.test-support.ts (modified, +1/-1)
• extensions/browser/src/cli/command-format.ts (modified, +1/-1)
• extensions/browser/src/config/config.ts (modified, +1/-1)
• extensions/browser/src/core-api.ts (modified, +25/-20)
• extensions/browser/src/doctor-browser.ts (modified, +1/-1)
• extensions/browser/src/gateway/auth.ts (modified, +1/-1)
• extensions/browser/src/gateway/startup-auth.ts (modified, +1/-1)
• extensions/browser/src/infra/errors.ts (modified, +1/-1)
• extensions/browser/src/infra/fs-safe.ts (modified, +1/-1)
• extensions/browser/src/infra/net/proxy-env.ts (modified, +1/-1)
• extensions/browser/src/infra/net/ssrf.ts (modified, +1/-1)
• extensions/browser/src/infra/path-guards.ts (modified, +1/-1)
• extensions/browser/src/infra/ports.ts (modified, +1/-1)

PR #61506: fix: prefer final-answer text across web chat surfaces

• Repository: openclaw/openclaw
• Author: steipete
• State: closed | merged: True
• Link: https://github.com/openclaw/openclaw/pull/61506

Description (problem / solution / changelog)

Summary

• prefer assistant final_answer text over commentary in web chat and previews
• suppress commentary text when completed WS replies are persisted
• share assistant phase helpers and patch the remaining chat.history phase-blind path

Verification

• pnpm test src/shared/chat-message-content.test.ts src/agents/pi-embedded-utils.test.ts src/agents/pi-embedded-subscribe.handlers.messages.test.ts src/agents/openai-ws-stream.test.ts src/gateway/session-utils.fs.test.ts ui/src/ui/chat/message-extract.test.ts
• pnpm build

Notes

• pnpm check on fresh origin/main is currently red from unrelated typing drift outside this branch's touched surface.
• follow-up to #59150 / #59643

Changed files

• CHANGELOG.md (modified, +1/-0)
• src/agents/openai-ws-message-conversion.ts (modified, +38/-48)
• src/agents/openai-ws-stream.test.ts (modified, +46/-17)
• src/agents/openai-ws-stream.ts (modified, +4/-15)
• src/agents/pi-embedded-subscribe.handlers.messages.ts (modified, +3/-43)
• src/agents/pi-embedded-utils.ts (modified, +5/-31)
• src/gateway/server-methods/chat.ts (modified, +2/-29)
• src/gateway/server.chat.gateway-server-chat.test.ts (modified, +23/-0)
• src/gateway/session-utils.fs.test.ts (modified, +50/-0)
• src/gateway/session-utils.fs.ts (modified, +10/-0)
• src/shared/chat-message-content.test.ts (modified, +111/-1)
• src/shared/chat-message-content.ts (modified, +146/-0)
• ui/src/ui/chat/message-extract.test.ts (modified, +35/-0)
• ui/src/ui/chat/message-extract.ts (modified, +3/-1)