[eslint-refiner] ESLint Refiner — Daily Report (2026-10-04)

[github-actions[bot]]

Overview

Daily refinement pass over eslint-factory's 64 custom TypeScript ESLint rules. HEAD (d952bc2, #65492) only touched .github/scripts/aw_issue_clustering*.py — out of scope — so today's review fell back to re-verifying rules whose issues had just expired, plus closing out a follow-up lead from yesterday's Codex-engine PR audit.

Key metrics

• Rule count: 64 (unchanged)
• Issues filed: 1
• Rules re-verified: 3 (no-math-minmax-array-spread, require-escaped-regexp-interpolation, no-unsafe-catch-error-property)
• Open eslint-factory issues: 2 (no-github-request-interpolated-route: fallback-expression Octokit client aliases (x || github, ternary) bypass client detection #64186, no-unsafe-catch-error-property: rename alias of catch param bypasses unsafe-property detection #65255) + 1 filed today

Highlights

• Confirmed fixed, theme retired: no-math-minmax-array-spread's identifier-initializer bounded-array false positive (previously no-math-minmax-array-spread: false positive on identifiers holding statically-bounded arrays #63781) is now resolved in source — isUnboundedSpreadArgument traces a const identifier back through an ArrayExpression (or chained .filter()/.slice()) initializer, and the rule's own test suite covers the exact live shape from daily_aic_api_budget.cjs. No refiling needed.
• Still open, refiled: require-escaped-regexp-interpolation has no way to recognize an intentional regex-fragment constant (e.g. a character class meant for composition into a larger pattern). The gap is unchanged in source since its first filing (require-escaped-regexp-interpolation: no recognition for regex-fragment constants (5 sites already need manual disable, 1 unsupp #63566, expired unfixed). Refiled with the same 6 live sites, one of which (handle_agent_failure.cjs:72) is still an unsuppressed live warning.
• Likely already fixed, not refiled: yesterday's no-unsafe-catch-error-property rename-alias finding (no-unsafe-catch-error-property: rename alias of catch param bypasses unsafe-property detection #65255) appears to already be resolved in current source — caughtName() now performs one-hop const-alias resolution. Left open to confirm on its own expiry rather than risk a false refiling.

Full per-rule detail

no-math-minmax-array-spread — Re-read eslint-factory/src/rules/no-math-minmax-array-spread.ts. isBoundedArrayExpression() (new since the 09-27 finding) recognizes a const bound to an array literal with no spread elements, or such a literal passed through .filter()/.slice(). The rule's test file (no-math-minmax-array-spread.test.ts:57-67) explicitly asserts const times = [reset, retryAt].filter(value => Number.isFinite(value)); Math.max(...times); is valid — the exact shape at daily_aic_api_budget.cjs:24-25. Theme retired.

require-escaped-regexp-interpolation — isLiteralSafeForRegexp still rejects any const string literal containing a regex metacharacter, with no opt-in for a deliberately-authored fragment. Live sites unchanged: sanitize_content_core.cjs (URL_START_DELIMITERS, URL_AUTHORITY_CHAR — 3 interpolation sites), model_aliases.cjs (VERSION_SUFFIX_PATTERN), glob_pattern_helpers.cjs (regexPattern) — all 5 already carry eslint-disable-next-line suppressions — and handle_agent_failure.cjs:72 (AWF_API_PROXY_HOST_RE_SOURCE), still unsuppressed. Filed as a recurrence of #63566; this is the theme's 2nd filing, not yet at the 2-expiry retirement threshold.

no-unsafe-catch-error-property — Re-read the rule behind #65255 (filed 2026-10-03 for a rename-alias blind spot at codex_harness.cjs:328,571,733). caughtName() (lines 207-218) now resolves a const alias of the catch parameter one hop back, and catchParamUnchangedSinceAlias() (lines 220-229) guards against the alias being reassigned before use. Tracing codex_harness.cjs:328 (const err = /** @type {Error} */ error; then err.message) through the current logic, the rule would now correctly flag it. No new issue filed; next run will confirm whether #65255 closes as a genuine fix.

Other checks, no issue: process_runner.cjs:249-250 has an unguarded error.code/error.message access inside a catch with no suppression — a true positive no-unsafe-catch-error-property already flags correctly (not a rule gap). codex_session.cjs/unified_session_payload.cjs's structuredClone usage (9 call sites) and module-level constants were checked against prefer-structured-clone and no-duplicate-constant-values respectively — both clean, closing out the remaining follow-up leads from the #65199 Codex-engine PR audit.

Next actions

• Watch #65255 at its ~2026-10-10 expiry — source read suggests it's already fixed; only refile if a genuinely new alias shape (2-hop rename, destructuring) turns up.
• Watch today's require-escaped-regexp-interpolation recurrence at its ~2026-10-11 expiry — a 2nd unfixed expiry would cross the 2-expiry retirement threshold for this theme.
• #64186 (no-github-request-interpolated-route fallback-expression alias gap) is now 5 days old, approaching the typical first-expiry window.
• Chronic, unactionable in this workflow (no create_pull_request safe-output): the shared try-catch-rule-utils.ts VariableDeclaration-suggestion gap affecting ~14 rules remains twice-filed-and-expired (eslint-factory: shared try/catch suggestion builder silently skips VariableDeclaration call sites #57868, eslint-factory: shared try/catch suggestion builder still skips VariableDeclaration call sites -- recurrence of expired issue 57 #59891) — tracked here for visibility only, will not be refiled a third time.

Generated by 🤖 ESLint Refiner · claude · agent · 240.6 AIC · ⌖ 6.05 AIC · ⊞ 5.1K · ◷

• expires on Oct 4, 2026, 11:11 PM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-10-05T07:11:35.610Z.

Closed by Workflow