Best possible solution:

Keep this in the normal maintainer PR flow, with final merge gated on the remaining CI and any maintainer review for the protected label.

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

Not applicable; this is a refactor/private API cleanup PR, not a bug report with a failing user flow. The behavior-preservation claim is source-checkable through the accessor and model override paths plus focused tests.

Is this the best way to solve the issue?

Yes; this is the best current shape for this slice because it reuses the existing session accessor seam, preserves Feishu latest-read semantics, and preserves Mattermost parent model override inheritance without widening private runtime barrels.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 4ae0a5d958a0.

Label changes:

• add P3: This is a low-risk cleanup/refactor slice with no intended user-visible behavior change.
• 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: The PR is maintainer-labeled private API/barrel cleanup with focused source and unit proof; the external contributor real-behavior proof gate does not apply here.

Label justifications:

• P3: This is a low-risk cleanup/refactor slice with no intended user-visible behavior change.
• 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: The PR is maintainer-labeled private API/barrel cleanup with focused source and unit proof; the external contributor real-behavior proof gate does not apply here.

PR surface:

Source +38, Tests +75, Other +2. Total +115 across 11 files.

What I checked:

• Root and scoped policy read: Read root AGENTS.md plus extensions/AGENTS.md, src/plugin-sdk/AGENTS.md, scripts/AGENTS.md, and test/AGENTS.md; plugin/SDK boundary guidance drove the review of private barrels, SDK subpath changes, and boundary guard coverage. (AGENTS.md:1, 4ae0a5d958a0)
• Full PR diff inspected: The patch removes Feishu/Mattermost private-barrel whole-store exports, migrates the two runtime read paths to targeted accessors, and adds focused tests for Feishu latest reads and Mattermost parent override inheritance. (extensions/mattermost/src/mattermost/model-picker.ts:236, 10b26fd3850e)
• Accessor contract supports latest reads: Current main already has SessionAccessScope.readConsistency?: "latest" and loadSessionEntry bypasses cache when latest consistency is requested; the PR threads that existing internal contract through the SDK runtime adapter. (src/config/sessions/session-accessor.ts:907, 4ae0a5d958a0)
• Mattermost parent inheritance compared against shared helper: resolveStoredModelOverride already accepts a parent session key plus a store map; the PR preserves that behavior by targeted-reading the parent entry and passing a one-entry map rather than exposing the whole store to the UI path. (src/auto-reply/reply/stored-model-override.ts:36, 4ae0a5d958a0)
• Related tracker context: The linked tracker explicitly lists this PR as the current 3.1b bundled Feishu/Mattermost runtime-barrel cleanup and says live proof is not required for this private API/barrel slice.
• CI and local evidence checked: The PR body reports focused Vitest, boundary guard, git diff --check, and branch autoreview; live check data showed check-session-accessor-boundary, check-prod-types, check-lint, check-test-types, and security scans succeeded while some broad CI shards were still running. (10b26fd3850e)

Likely related people:

• jalehman: Recent merged history on src/plugin-sdk/session-store-runtime.ts and the linked tracker shows repeated Path 3 session-accessor migration slices and boundary guard updates. (role: recent session-accessor stack contributor; confidence: high; commits: b58e6e0734f2, acc2a0ee7297, c588606a9bf3; files: src/plugin-sdk/session-store-runtime.ts, scripts/check-session-accessor-boundary.mjs, test/scripts/check-session-accessor-boundary.test.ts)
• steipete: History on the Mattermost model-picker and plugin SDK session helper surfaces includes model catalog extraction, Mattermost picker maintenance, SDK surface reduction, and session entry workflow helpers. (role: adjacent plugin SDK and Mattermost area contributor; confidence: medium; commits: 30e1556cdac9, c4a5bba800d7, 827b0de0ce74; files: extensions/mattermost/src/mattermost/model-picker.ts, src/plugin-sdk/session-store-runtime.ts)
• vincentkoc: History on Feishu reasoning preview and the session SQLite rollback path makes this a relevant routing candidate for Feishu behavior and migration safety. (role: adjacent Feishu/session migration contributor; confidence: medium; commits: d609f71c9b74, 7f1d82ab2518; files: extensions/feishu/src/reasoning-preview.ts, src/plugin-sdk/session-store-runtime.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.