fix(frontmatter,prompts): close two regressions from the Bun built-in swap#68
Merged
Merged
Conversation
… swap An independent Codex cross-model review of the yaml→Bun.YAML and @inquirer/confirm→node:readline migration (#65) surfaced two real Medium regressions, both reproduced and confirmed: - Bun.YAML silently keeps the last value on duplicate mapping keys, where the `yaml` package threw "Map keys must be unique". parseFrontmatterStrict now detects duplicate top-level keys itself, so lint:skills / lint:knowledge keep catching them. The scan is column-0 only — nested mappings, list items, and block-scalar continuations are always indented, so they never false-trip (covered by a test). - promptYn (readline) hung on Ctrl+C: readline's close() does not unblock a pending question() — only an AbortSignal rejects it. Wired a SIGINT→AbortController so Ctrl+C falls back to the default, restoring the behavior @inquirer/confirm gave for free. Full suite 618 pass / 0 fail; lint:skills/lint:knowledge clean on the real corpus.
arzafran
added a commit
that referenced
this pull request
Jun 22, 2026
Patch release covering the unreleased fixes since 11.28.0: - #68: Bun built-in swap follow-ups (Bun.YAML duplicate-key detection in parseFrontmatterStrict; promptYn Ctrl+C via SIGINT→AbortController) - install-summary count fix: skills/ counts SKILL.md subdirs (35 not 1), docs/ counts recursively (22), with regression tests Bumps VERSION, plugin.json, and prepends the 11.28.1 CHANGELOG section.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this does
A Codex cross-model review of the dependency swap (#65) caught two real behaviour regressions that our own tests missed — both reproduced, both fixed here. Neither is critical, but each quietly weakens something the old libraries handled for free.
Summary
Bun.YAML.parsesilently keeps the last value on a duplicate mapping key (name: foothenname: bar→{ name: "bar" }), where the oldyamlpackage threw "Map keys must be unique". That meanslint:skills/lint:knowledgestopped flagging duplicate keys in skill/agent frontmatter.parseFrontmatterStrictnow detects duplicate top-level keys itself and reports them. The scan is column-0 only — nested mappings, list items, and block-scalar continuations are always indented, so they never false-trip (there's a test for exactly that).promptYnno longer hangs on Ctrl+C. With@inquirer/confirm, SIGINT threw and the caller fell back to the default. With rawnode:readline, Ctrl+C while the question is pending did nothing —rl.close()does not unblock a pendingquestion(); only anAbortSignalrejects it (verified directly in Bun). ASIGINT → AbortControllernow restores the fall-back-to-default behaviour so interactive installer/merge prompts can't wedge.Test Plan
bun test— 618 pass, 0 fail (+2 frontmatter tests: duplicate-key flagged, indented repeats not false-flagged)bun run lint:skills/lint:knowledgeclean on the real corpus (no false positives from the new detector)bun run typecheckclean ·bun run lintexit 0rl.question(prompt, { signal })rejects withAbortErroron abort;rl.close()alone leaves it hung. (The TTY+signal path isn't unit-testable deterministically, so it's verified by the mechanism probe, not a unit test.)