🤖 refactor: auto-cleanup

[mux-bot]

Summary

Long-lived auto-cleanup PR maintained by the Auto-Cleanup Agent. Each pass lands at most one extremely low-risk, behavior-preserving cleanup drawn from recently merged main activity, then advances the checkpoint below.

This pass

Deduplicated two repeated fallback error strings in src/browser/components/AutomationModal/AutomationModal.tsx, introduced by #3560 ("centralize workflow automations"). The literals "Failed to save automation." and "Failed to remove automation." each appeared three times across the save/remove flows (pre-check, RPC-result, and thrown-error/catch paths). They are now SAVE_AUTOMATION_ERROR_MESSAGE and REMOVE_AUTOMATION_ERROR_MESSAGE module-level constants, so the wording stays consistent if it ever changes.

• Behavior-preserving: each constant holds the exact same string previously inlined at every call site; the rendered/thrown text is unchanged. No test asserts these literals, and the 13 AutomationModal.test.tsx tests still pass.
• Scoped: single file, no new imports, no exported-API changes.

Validation

• bun test src/browser/components/AutomationModal/AutomationModal.test.tsx — 13 pass / 0 fail.
• make static-check — eslint, tsc (both configs), and Prettier pass. The only failing step is fmt-shell-check, which requires shfmt (not installed in this environment); no shell scripts were touched, so CI (which has shfmt) is unaffected.

Prior passes

• Replaced a duplicated "wfr_" literal in src/node/services/workflows/WorkflowRunner.ts with the canonical isWorkflowRunTaskId() predicate from src/node/services/tools/taskId.ts (after 🤖 feat: add first-class nested workflows #3565). Behavior-preserving inside the Array.find() predicate (both forms are falsy for a missing taskId and perform the same startsWith check for a present one).
• Removed a duplicated "chat.jsonl" literal in src/node/services/analytics/etl.ts by re-exporting CHAT_FILE_NAME from the canonical src/common/constants/paths.ts (after 🤖 perf: rotate sealed chat history out of chat.jsonl at compaction boundaries #3541).
• Extracted a harvestRecordTime helper in src/node/services/memoryConsolidationService.ts so findNewestHarvestRecord and pruneHarvestRecords derive the completedAt ?? startedAt ordering key from one place (after 🤖 feat: add memory harvest pipeline #3558).

Auto-cleanup checkpoint: f96aa91

---

Generated with mux • Model: anthropic:claude-opus-4-8 • Thinking: xhigh • Cost: n/a

[mux-bot]

@codex review

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.

[mux-bot]

⚠️ Auto-fixup analysis: the CI failure was a flaky test, not caused by this PR's changes. No fix pushed.

Failure: Run 27552085157 (commit 141167eb8) failed only in Test / Unit, on:
built-in deep-research workflow > keeps prototype-key source identifiers during dedupe
with a Bun engine-level crash: RuntimeError: Unreachable code should not be executed.

Why it's flaky / unrelated:

• This PR's only diff is a pure refactor of src/node/services/memoryConsolidationService.ts (extracting a harvestRecordTime helper). It has no connection to the deep-research workflow source-dedupe path that crashed.
• RuntimeError: Unreachable code should not be executed is an intermittent Bun JIT/engine crash, not a deterministic assertion failure.
• Reproduced locally on the PR head: the originally-failing test passed in 3/3 full-file runs, while a different test (auto-fix checkpoint retry preserves completed patch after HEAD advances) flaked once — confirming non-determinism in this file.
• memoryConsolidationService.test.ts: 36/36 pass.
• A subsequent CI run on the current head a913e8d48 (run 27552691244) already completed successfully.

Action: None required for this PR. The flakiness lives in builtInWorkflowDefinitions.test.ts and should be tracked/stabilized separately.

---

Generated with mux • Model: anthropic:claude-opus-4-8 • Thinking: xhigh

[mux-bot]

@codex review

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.

[mux-bot]

Auto-cleanup pass note: the one failing check (Test / Unit) is pre-existing on main, not caused by this change.

• This pass only touches src/node/services/analytics/etl.ts (re-export the canonical CHAT_FILE_NAME instead of redefining the "chat.jsonl" literal). historyService.ts/historyService.test.ts are byte-identical to origin/main.
• The failing tests are HistoryService > getMessagesForCompactionEpoch > … (introduced by 🤖 perf: rotate sealed chat history out of chat.jsonl at compaction boundaries #3541). Test / Unit is currently red on main HEAD (a6d11e3b5) and the prior commit (a8f811f96) independently of this PR.
• All other checks pass here, including CI Static Checks (full make static-check incl. shfmt), Test / Integration, builds, E2E, Storybook, and Chromatic.

Per the auto-cleanup agent's "extremely low-risk, behavior-preserving" constraint, this pass does not attempt to fix the unrelated HistoryService test regression. Pausing for human direction on that pre-existing failure.

[mux-bot]

⚠️ Auto-fixup could not resolve CI failure: the failing tests are not caused by the cleanup commits — they are pre-existing failures already red on main. Manual intervention needed.

What failed

Test / Unit failed with 10 tests (failed run 27564995706):

• HistoryService > getMessagesForCompactionEpoch (3 tests)
• MemoryConsolidationService (7 tests, incl. 2 timeouts)

Why it is not from this branch

The same 10 tests fail on main HEAD's own CI (acb98ed6c, run 27564747006), and I reproduced them locally at the merge base a6d11e3b5 — neither of which contains this branch's two commits. Both cleanup commits are behavior-preserving and untouched by the affected code:

• 426bc8d41 only refactors memoryConsolidationService.ts ordering (identical values).
• 2df82f753 only re-exports CHAT_FILE_NAME in analytics/etl.ts.

Root cause (pre-existing, from #3558)

src/node/services/historyService.ts getMessagesForCompactionEpoch calls this.iterateForward(workspaceId, …), but iterateForward(filePath, …) expects a file path (every other call site passes getChatHistoryPath(...) / getChatArchivePath(...)). With a bare workspaceId, fs.stat hits ENOENT, the visitor never runs, the summary is never found, and the call returns Err("Compaction summary not found"). The MemoryConsolidationService harvest tests fail/time out downstream because the memory harvest reads compaction-epoch messages through this broken path. The line was introduced by 5584b6437 (#3558), which is on main.

Action

Per the auto-fixup scope ("only fix what the cleanup broke; do not add unrelated changes"), I did not push a fix. The fix belongs on main (e.g., pass this.getChatHistoryPath(workspaceId) and iterate the sealed archive too). Once main is green, this PR's CI should pass on rebase/merge.

[mux-bot]

@codex review

This pass replaces a duplicated "wfr_" literal in WorkflowRunner.ts's nested-step drift check with the canonical isWorkflowRunTaskId() predicate from taskId.ts. Behavior-preserving (the result is only used as a boolean inside Array.find(); both forms are falsy for a missing taskId).

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.

[mux-bot]

@codex review

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.