Best possible solution:

Land the Android-side visible-role contract after required CI and maintainer/security review, while keeping gateway projection behavior for richer clients unless a separate cross-client policy change is approved.

Do we have a high-confidence way to reproduce the issue?

Yes. Source inspection on current main shows a chat.history row with a non-visible role and text content is parsed, cached, and rendered as an OpenClaw chat bubble.

Is this the best way to solve the issue?

Yes. The PR fixes the Android client boundary at live history parsing, offline cache read/write, and final rendering while avoiding a global gateway deletion that could regress richer history clients.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 827402243d7f.

Label changes:

• add P1: The PR targets a current Android chat leak of internal reasoning/tool details that can affect real users and is linked to an open security-impact issue.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: Not applicable as a contributor proof gate because the PR is maintainer-labeled; the body includes Testbox seam-test evidence but no physical Android screenshot or recording.

Label justifications:

• P1: The PR targets a current Android chat leak of internal reasoning/tool details that can affect real users and is linked to an open security-impact issue.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: Not applicable as a contributor proof gate because the PR is maintainer-labeled; the body includes Testbox seam-test evidence but no physical Android screenshot or recording.

PR surface:

Docs +2, Other +75. Total +77 across 8 files.

What I checked:

• Root policy read: Read the full root AGENTS.md and the Android scoped AGENTS.md; root protected-label, whole-surface review, security-sensitive bug, and proof guidance affected the review, while Android scoped guidance was release/license-specific. (AGENTS.md:1, 827402243d7f)
• Current main still accepts arbitrary history roles: On current main, parseHistory copies any incoming non-empty role from chat.history into a ChatMessage, so text-bearing toolResult or internal rows can enter Android state before this PR. (apps/android/app/src/main/java/ai/openclaw/app/chat/ChatController.kt:1648, 827402243d7f)
• Current main already drops typed non-text blocks: The existing parser maps only text/input_text/output_text and image blocks, returning null for other content block types; this supports the PR body's claim that the remaining leak is row-role visibility rather than typed thinking-block parsing alone. (apps/android/app/src/main/java/ai/openclaw/app/chat/ChatController.kt:1814, 827402243d7f)
• Current renderer displays unknown roles as assistant-like bubbles: Current ChatMessageBubble filters content to text/image but maps every non-user/non-system role to the OpenClaw label, which is the visible leak path the PR changes defensively. (apps/android/app/src/main/java/ai/openclaw/app/ui/chat/ChatMessageViews.kt:62, 827402243d7f)
• Current cache persists arbitrary roles: Current RoomChatTranscriptCache saves and reloads text rows with message.role/row.role unchanged, so a leaked internal row can also persist into the offline transcript cache before this PR. (apps/android/app/src/main/java/ai/openclaw/app/chat/ChatTranscriptCache.kt:237, 827402243d7f)
• PR diff covers all Android ingress/cache/render points: The PR normalizes visible roles at live-history parsing, cache load/save, and message rendering, and adds tests for live history and cache round-trip dropping internal roles while preserving custom and normalized assistant rows. (apps/android/app/src/main/java/ai/openclaw/app/chat/ChatModels.kt:1, 15201f2dfec2)

Likely related people:

• LeonidasLux: Current blame for the Android parser, content parser, message renderer, and transcript-cache rows points to merged commit 0f01930 from PR fix(plugins): discover config load.paths plugins when index has entries #99196. (role: introduced current Android chat surface; confidence: high; commits: 0f01930c8eeb; files: apps/android/app/src/main/java/ai/openclaw/app/chat/ChatController.kt, apps/android/app/src/main/java/ai/openclaw/app/ui/chat/ChatMessageViews.kt, apps/android/app/src/main/java/ai/openclaw/app/chat/ChatTranscriptCache.kt)
• vincentkoc: Live PR metadata shows vincentkoc merged PR fix(plugins): discover config load.paths plugins when index has entries #99196, whose commit currently owns the affected Android chat lines. (role: merger/reviewer of current Android file shape; confidence: medium; commits: 0f01930c8eeb; files: apps/android/app/src/main/java/ai/openclaw/app/chat/ChatController.kt, apps/android/app/src/main/java/ai/openclaw/app/ui/chat/ChatMessageViews.kt, apps/android/app/src/main/java/ai/openclaw/app/chat/ChatTranscriptCache.kt)
• steipete: Commit b4b1984 recently touched Android ChatController for managed-worktree chat sessions, and this person owns the current PR plus nearby gateway/chat-history work. (role: recent adjacent chat contributor; confidence: medium; commits: b4b1984fe7a9, 15201f2dfec2; files: apps/android/app/src/main/java/ai/openclaw/app/chat/ChatController.kt, src/gateway/server-methods/chat.ts, src/gateway/chat-display-projection.ts)

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.