[eslint-refiner] ESLint Refiner — Daily Report — 2026-08-14

[github-actions[bot]]

Overview

Reviewed 3 previously-unaudited eslint-factory rules against live actions/setup/js/** code and filed 3 grounded refinement issues. Repo-memory was significantly stale (last recorded run 2026-07-08, 12 rules) — eslint-factory/src/index.ts now registers 43 rules, and a search of closed issues surfaced at least one full intermediate run (issues #51311-#51313, ~2026-08-08) that was never recorded in this memory. Memory has been rebuilt from ground truth this run.

Issues filed this run

1. require-error-code-in-thrown-error — regex \b word-boundary bug: an embedded code like SAFE_OUTPUT_E099 (constant name ends in a 3-digit code, preceded by _) never matches (be/redacted)[0-9]{3}\b/ because _ is a word character, so a correctly-coded throw is flagged as missing a code. Grounded at 5 live sites (dismiss_pull_request_review.cjs:249, add_labels.cjs:365, comment_memory.cjs:41, merge_pull_request.cjs:71,184). Distinct from the three prior closed issues on this rule (eslint-factory: require-error-code-in-thrown-error misses thrown custom Error subclasses #51311 subclasses, eslint-factory: require-error-code-in-thrown-error skips CallExpression message args (helper-built messages never audited) #51312 CallExpression args, eslint-factory: require-error-code-in-thrown-error false-positives on Identifier args that already carry a code #51313 Identifier resolution) — none cover this boundary case.
2. require-fetch-response-body-try-catch — message says an unwrapped body-read "will crash the action," but every entrypoint is required to have a top-level catch (enforced by require-async-entrypoint-catch), so it's never truly uncaught — just less contextualized. Grounded at artifact_client.cjs:237,292,295. Wording-only ask, no detection-logic change.
3. require-sync-exec-timeout — hasTimeoutOption accepts any non-nullish timeout value, including timeout: 0, which Node documents as "no timeout" (identical to omitting it). Currently ungrounded in the live corpus (no timeout: 0 call site found), but a real latent soundness gap — same class as the earlier catchless-tryfinally-protective-FN fix for require-fs-sync-try-catch.

Rules reviewed but not filed against (no new defect found)

• Re-confirmed require-async-entrypoint-catch-style backstop reasoning holds across push_signed_commits.cjs's two custom Error subclasses (both already carry ERR_VALIDATION-prefixed messages at every throw site — ruled out as a live gap for eslint-factory: require-error-code-in-thrown-error misses thrown custom Error subclasses #51311's class).
• require-sync-exec-timeout core binding-resolution (ESM import, CJS destructure, member-call, function-local require) verified correct against ~11 genuine missing-timeout violations across 5 files (start_mcp_gateway.cjs, merge_remote_agent_github_folder.cjs, apply_samples.cjs, safeoutputs_cli.cjs, get_current_branch.cjs) — the rule's detection is sound, only the timeout: 0 edge case above is a gap.

Repo-memory drift

Memory (state.json) had not been updated since 2026-07-08 (12 rules), but the registry has grown to 43 and at least one additional run (issues #51311-#51313, ~2026-08-08) happened without leaving a memory trace — likely a missed push_repo_memory on that run, or a memory-branch mismatch. This run's memory update includes the current rule count and an explicit backlog of the ~29 still-never-reviewed rules so future runs can pick up systematically instead of re-discovering the gap.

Next priorities

• Continue working through the ~29 unreviewed rules (see updated nextPriorities in repo-memory), prioritizing newest-registered first.
• Spot-check that push_repo_memory is actually landing after future runs — verify next run's state.json reflects today's update before assuming memory is reliable.

References:

• Prior related issues: eslint-factory: require-error-code-in-thrown-error misses thrown custom Error subclasses #51311, eslint-factory: require-error-code-in-thrown-error skips CallExpression message args (helper-built messages never audited) #51312, eslint-factory: require-error-code-in-thrown-error false-positives on Identifier args that already carry a code #51313 (same rule, different gaps, both closed/expired)

Generated by 🤖 ESLint Refiner · agent · 255.1 AIC · ⌖ 34.6 AIC · ⊞ 4.7K · ◷

• expires on Aug 14, 2026, 10:01 PM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-08-15T06:01:05.839Z.

Closed by Workflow