Best possible solution:

Land the focused payload fix after a credentialed Feishu/Lark streaming-card smoke test or an explicit maintainer proof override, then let the linked bug close through the merged PR.

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

Yes at source level: current main slices Feishu CardKit updates down to suffixes while the official streaming-card guide describes sending full text and letting the platform compute the visible increment. I did not run a live Feishu/Lark tenant reproduction in this read-only review.

Is this the best way to solve the issue?

Yes, this is the narrowest maintainable fix I found: the changed helper is the single payload decision point before updateCardContent, and the replace path for final rewrites remains separate. The remaining gap is live external proof, not a better code location.

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 4932391e8a78.

Label changes:

• add P1: The linked bug reports Feishu streaming cards showing only the newest suffix for affected users, breaking a real channel workflow when streaming is enabled.
• add merge-risk: 🚨 message-delivery: The PR changes the text payload sent to Feishu CardKit, so incorrect behavior would directly affect delivered chat-card content.
• add rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🐚 platinum hermit.
• add status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body provides a focused Vitest run but explicitly says no live Feishu/Lark tenant flow was tested; redacted screenshot, recording, logs, or terminal/live output from a real tenant should be added before merge and updating the PR body should trigger rereview. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Label justifications:

• P1: The linked bug reports Feishu streaming cards showing only the newest suffix for affected users, breaking a real channel workflow when streaming is enabled.
• merge-risk: 🚨 message-delivery: The PR changes the text payload sent to Feishu CardKit, so incorrect behavior would directly affect delivered chat-card content.
• rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🐚 platinum hermit.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body provides a focused Vitest run but explicitly says no live Feishu/Lark tenant flow was tested; redacted screenshot, recording, logs, or terminal/live output from a real tenant should be added before merge and updating the PR body should trigger rereview. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

PR surface:

Source -1, Tests 0. Total -1 across 2 files.

What I checked:

• Current main sends suffix-only content: On current main, resolveStreamingCardAppendContent returns nextText.slice(previousText.length) when the new text starts with the previously sent text, and updateCardContent passes that string as the CardKit content payload. (extensions/feishu/src/streaming-card.ts:157, 4932391e8a78)
• CardKit update call boundary: The Feishu streaming session sends PUT /cardkit/v1/cards/{cardId}/elements/content/content with JSON body { content, sequence, uuid }, so the helper output is exactly the external CardKit text update payload. (extensions/feishu/src/streaming-card.ts:350, 4932391e8a78)
• Patch changes the payload mode: The PR patch removes the suffix-slicing branch and returns nextText after the unchanged/empty guard, with tests updated to expect hello small, hello!, hello world, and hello world! as full payloads. (extensions/feishu/src/streaming-card.ts:158, 1c4a18da734f)
• Existing focused tests prove the current mismatch: The base tests currently assert suffix payloads for throttled, natural-boundary, failed-update, and non-OK retry paths; the PR changes those same cases to full-content payloads without changing unrelated flow control. (extensions/feishu/src/streaming-card.test.ts:181, 4932391e8a78)
• External CardKit contract check: The official Feishu streaming-card guide says streaming text updates pass full text content, and if old text is a prefix of the new text the platform computes and renders the added suffix itself.
• Linked bug remains paired with this PR: The PR body says it closes [Bug]: Feishu streaming card: only the last character displayed after streaming completes #88867, whose supplied context reports 100% reproduction for Feishu streaming cards showing only the last suffix while chat-list preview has the full text. (1c4a18da734f)

Likely related people:

• hclsys: GitHub commit history shows f436b43 as the Feishu CardKit text-delta change that introduced the suffix-oriented streaming path being corrected here. (role: introduced behavior; confidence: high; commits: f436b4310a0c; files: extensions/feishu/src/streaming-card.ts)
• qiangu: GitHub commit history shows 3b8ab4e recently moved ordinary Feishu streaming replies through CardKit cards, making this streaming-card path broadly visible. (role: recent area contributor; confidence: high; commits: 3b8ab4e11232; files: extensions/feishu/src/streaming-card.ts, extensions/feishu/src/reply-dispatcher.test.ts)
• ArthurNie: GitHub commit history shows 7c15c27 recently changed Feishu visible-reply fallback behavior around streaming closeout and delivery state. (role: recent adjacent contributor; confidence: medium; commits: 7c15c2765ebb; files: extensions/feishu/src/streaming-card.ts, extensions/feishu/src/reply-dispatcher.test.ts)
• steipete: Local and GitHub history show repeated recent Feishu streaming-card and test maintenance, including token expiry, helper export trimming, and lint/readability updates in this area. (role: recent area contributor; confidence: medium; commits: 04de01f8cfe9, b388209eaf72, 27dde7a4d69b; files: extensions/feishu/src/streaming-card.ts, extensions/feishu/src/reply-dispatcher.test.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.