[eslint-refiner] ESLint Refiner — Daily Report 2026-07-03

[github-actions[bot]]

Overview

Focus this run: first analysis of the newly added require-async-entrypoint-catch rule, plus re-verification of prior open issues. The factory grew from 6 → 7 rules since the last run.

Key metrics

Metric	Value
Rules in factory	7
New rule analyzed	require-async-entrypoint-catch
Issues filed this run	2 (both new-rule; limit 3)
Prior open issues re-verified still valid	9
Prior issues refiled	0 (no duplicates created)

Issues filed

Both target the newest rule, whose detection is currently the narrowest of the family:

1. FN — declaration-shape blindness: rule only tracks async FunctionDeclarations and bare Identifier() statements, so const main = async () => {}; main(); and main().then(...) without .catch() escape it.
2. FP — scope-unaware matching: call sites are matched by bare name against a module-wide Set<string>, so a shadowed same-named sync function is misflagged as an unhandled async rejection (.catch(...) suggestion would be wrong at runtime).

Live grounding

Corpus check over actions/setup/js/**/*.cjs

• True positives the rule already catches: parse_mcp_scripts_logs.cjs:415 and parse_token_usage.cjs:238 — async function main() invoked bare under if (require.main === module). The rule works well for the dominant idiom.
• Correctly ignored: sync function main() entrypoints (e.g. convert_gateway_config_copilot.cjs:97, evaluate_outcomes.cjs:1934).
• Arrow/expression async entrypoints: none are invoked bare today — the arrow-async helpers found (runMain, replayMain, runHandlerManager, substitutePlaceholders) are all await-guarded inside async wrappers. So issue rejig docs #1 is a preventive robustness/parity refinement, not a live defect — mirroring how the sibling error-property rules were hardened.

Prior open issues — still valid, deliberately not refiled

Re-verified against current source

• eslint-factory: no-core-setoutput-non-string misses negative/unary numeric literals (e.g. -1, +5) #42680 — no-core-setoutput-non-string still misses unary/negative numeric literals (nonStringKind matches Literal, not UnaryExpression). Confirmed.
• eslint-factory: no-core-setoutput-non-string only recognizes .length; misses .number/.size (live FN in close_entity_helpers.cjs) #42682 — same rule still recognizes only .length, not .number/.size. Confirmed.
• eslint-factory: no-unsafe-catch-error-property guard is whole-block, causing branch-order false negatives #42189 / eslint-factory: no-unsafe-promise-catch-error-property guard is whole-callback, causing branch-order false negatives #42915 — whole-block / whole-callback branch-order false negatives in the two no-unsafe-*-catch-error-property rules. Still present.
• eslint-factory: broaden no-unsafe-catch-error-property beyond message/stack/code to .status/.cause/.name #42188 / eslint-factory: no-unsafe-catch-error-property — add suggestions for stack/code and cover computed access #42190 — broaden catch-error-property coverage + suggestions/computed access. Open.
• eslint-factory: add a test suite for require-parseInt-radix (currently zero coverage) #42015 / eslint-factory: add a suggestion fix + fix broken docs URL for require-parseInt-radix #42016 / eslint-factory: require-parseInt-radix misses computed/indirect parseInt access #42017 / eslint-factory: require-json-parse-try-catch misses indirect/computed JSON.parse access #41962 / eslint-factory: try-block ancestry check is unsound for deferred callbacks #41963 — parseInt/json coverage, docs, and deferred-callback soundness. Open (alias/computed cases remain out of scope by design).

Close-recommendation note: this workflow cannot close issues (safe outputs are limited to create/report only). Maintainers may want to sweep the parseInt/json set — several ask for alias/computed handling that the rules intentionally exclude to keep false-positive risk low.

Next actions

• Consider a shared reporting helper between the two no-unsafe-*-catch-error-property rules to stop the recurring drift between them (maintainability).
• Explore a genuinely new recurring defect class beyond the error-handling family — e.g. core.setFailed non-string arguments (mirrors no-core-setoutput-non-string). Missing-await classes are attractive but hard to make low-false-positive without type information.

References: §28642396258

Generated by 🤖 ESLint Refiner · 147.1 AIC · ⌖ 12.9 AIC · ⊞ 4.7K · ◷

• expires on Jul 3, 2026, 10:27 PM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by ESLint Refiner.

A newer discussion is available at Discussion #43327.