Best possible solution:

Land the test-only reshuffle after remaining required checks finish, keeping gateway startup tests narrow and plugin-owned session-store discovery covered in infra and registry tests.

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

No deterministic current-main reproduction was run. The linked Actions job failed the compact shard and current source shows the gateway test exercises real plugin discovery, so the flake path is plausible but not a high-confidence repeated repro.

Is this the best way to solve the issue?

Yes. Moving plugin-owned migration behavior coverage into infra and registry tests is a narrower maintainable fix than keeping a gateway wrapper test that loads bundled plugin discovery.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 119dc4bd821e.

Label justifications:

• P3: This is low-risk test stabilization with no runtime, config, migration, dependency, or workflow file changes.
• 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: this is a maintainer-labeled test-only PR, and the PR body lists targeted test and autoreview validation instead of user-facing runtime proof.

PR surface:

Tests +44. Total +44 across 3 files.

What I checked:

• Protected maintainer signal: Live PR metadata shows the maintainer label and MEMBER author association, which means this cleanup workflow must keep the PR open for explicit maintainer handling. (9d85943ff0a6)
• Current-main broad gateway test: Current main still has a direct gateway startup session migration test that creates a Voice Call plugin config and runs the real startup migration path for plugin-owned agent discovery. (src/gateway/server-startup-session-migration.test.ts:32, 119dc4bd821e)
• Scoped gateway policy: The gateway scoped guide says startup paths and tests should not materialize bundled plugin runtime when static plugin-owned descriptors are enough, which supports narrowing this gateway wrapper test. (src/gateway/AGENTS.md:3, 119dc4bd821e)
• PR-head wrapper tests: On the PR head, the gateway startup migration test keeps coverage focused on logging, warnings, no-op behavior, and caught errors through injected migration deps. (src/gateway/server-startup-session-migration.test.ts:29, 9d85943ff0a6)
• PR-head infra discovery guard: The orphan-key migration test now mocks only listPluginDoctorSessionStoreAgentIds, verifies the config/env/pluginIds call, and proves the Voice Call session key is canonicalized. (src/infra/state-migrations.orphan-keys.test.ts:278, 9d85943ff0a6)
• Runtime migration contract: The production migration already accepts explicit additional agent IDs and otherwise resolves plugin-owned session-store agent IDs through doctor contracts, so the PR is changing test placement rather than runtime behavior. (src/infra/state-migrations.ts:5128, 119dc4bd821e)

Likely related people:

• obviyus: Merged the original session-key canonicalization work and later authored the PR that added orphaned session key migration and gateway startup migration coverage. (role: original migration owner and merger; confidence: high; commits: d3deacba8436, 56c079eba725, b3c69b941e6d; files: src/infra/state-migrations.ts, src/gateway/server-startup-session-migration.ts, src/gateway/server-startup-session-migration.test.ts)
• jzakirov: Authored the recent merged session migration performance PR that modified orphan-key migration tests and relied on startup migration as the canonicalization boundary. (role: recent area contributor; confidence: medium; commits: ed261c6720af, a86ca4f4ba97; files: src/infra/state-migrations.orphan-keys.test.ts, src/gateway/session-utils.ts, src/config/sessions/store.ts)
• jalehman: Authored the startup migration logging split in the recent session migration refactor PR and merged that adjacent area work. (role: recent adjacent contributor and merger; confidence: medium; commits: 42c922460a1f, a86ca4f4ba97; files: src/config/sessions/startup-migration.ts, src/gateway/server-startup-session-migration.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.