[eslint-refiner] ESLint Refiner Daily Report 2026-10-01

[github-actions[bot]]

ESLint Refiner — Daily Report (2026-10-01)

Overview: Quiet, disciplined day. No new rule shipped (ruleCount steady at 63) and today's HEAD commit (#64639) is genuinely docs-only despite touching actions/setup/js paths in its diffstat. No issues filed — the rotation candidates carried from prior runs all came up clean on a second pass. The standout item is a process win: yesterday's filed issue was fixed and correctly closed within 9 hours.

Key metrics

Metric	Value
Rules reviewed today	4 (1 confirmed-fixed re-check + 3 rotation re-checks)
New issues filed	0
Open eslint-factory issues	4 (#64186, #63781, #63566, #63359)
Oldest open issue	#63359, 6 days — approaching typical ~1wk first-expiry window
Total rules in eslint-factory	63 (unchanged)

Process win: clean fix + closure cycle

#64416 (no-empty-catch-block, filed 2026-09-30, flagged that the rule's keyword allowlist missed "must not (consequence)" rationale comments) was fixed and closed as completed within 9 hours. Verified live in eslint-factory/src/rules/no-empty-catch-block.ts: a new consequenceNotAllowedCommentRe pattern plus an extended negation regex now recognize that phrasing, matching the filed issue's proposal almost exactly.

This breaks a pattern seen twice before — #61044 and #61543 were also genuine upstream fixes, but both got bot-closed as not_planned instead of completed (independently re-confirmed today via the GitHub API). No action is available from this workflow (no close_issue/add_comment safe-output), but it's a positive signal worth a human's attention.

Rotation re-check detail (no new findings)

Three rules untouched since 2026-09-11 got a full second pass against the current corpus — all clean:

• no-caught-error-interpolation: checked the += assignment idiom (moot — the TemplateLiteral visitor already catches the nested template regardless of the enclosing operator), ternary-wrapped bare-identifier interpolation (${cond ? err : ...} — zero live occurrences; every live ternary branch already uses .message/.stack/getErrorMessage()/String(), never a bare identifier), and two-hop const alias chains — zero live occurrences.
• prefer-structured-clone: 2 live non-test call sites (pick_experiment.cjs:268, ledger_store.cjs:352/359) — both clone genuinely plain-JSON data (experiment run records, ledger records/metadata) with no function-valued properties anywhere in their files, so the autofix suggestion is correct with zero false-positive risk.
• require-lastindex-reset-before-global-exec-loop: audited all 9 live .exec()-in-while-loop regex sites across the corpus (copilot_harness.cjs, comment_memory_helpers.cjs, safe_outputs_config.cjs, patch_path_helpers.cjs, frontmatter_hash_pure.cjs ×3, log_parser_format.cjs, resolve_mentions.cjs) — every regex is declared with a function-local const pattern = /.../g, never at module scope, so each call gets a fresh regex and the rule correctly does not flag them.

This rotation list is now demoted to lowest-priority re-check cadence until a new feature commit touches error-handling or regex/.exec() code.

Open issue tracker snapshot

Issue	Rule	Age	Note
#64186	no-github-request-interpolated-route	3 days	fallback-expression Octokit client alias bypass
#63781	no-math-minmax-array-spread	4 days	identifier-initializer false positive
#63566	require-escaped-regexp-interpolation	5 days	regex-fragment-constant recognition gap
#63359	require-fetch-response-body-try-catch	6 days	chain-unwrap gap, recurrence of #61542 — watch for expiry next run

Next actions

• Watch #63359 for expiry next run — if it lapses unfixed, this would be its 2nd expiry (after #61542), crossing the established retire-after-2-expiries threshold.
• Primary detection method remains: check for a ruleCount bump or a verified (file-list-confirmed, not just diffstat) new-feature commit first; only fall back to rotation re-checks when both are flat, as today.
• Standing ask for a human/PR-capable workflow: the chronic try-catch-rule-utils.ts VariableDeclaration-suggestion gap (affects ~14 fs-sync/child-process rules, twice filed and expired as #57868/#59891) still can't be fixed from this workflow (no create_pull_request safe-output).

Generated by 🤖 ESLint Refiner · claude · agent · 228 AIC · ⌖ 5.74 AIC · ⊞ 5.2K · ◷

• expires on Oct 1, 2026, 9:33 PM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by ESLint Refiner.

A newer discussion is available at Discussion #64923.