release: prepare setup-benchmark-go-action v1.0.5

[cpunion]

Summary

• release the default-branch publishing and permission changes from feat: publish main benchmarks from the current workflow #6
• release signed unit deltas and paired comparison support from feat: report paired benchmark differences #7
• bump the package version to v1.0.5

Validation

• npm run check (37 tests)
• npm run build
• npm audit --omit=dev

[github-actions]

setup-benchmark-go-action

455431f331b8 | workflow run | long-term charts

Action bundle sizes

Bundle	Size	vs main
RecordBundle	330118 B	+0.0%
RenderBundle	330661 B	+0.0%

Parser and validation operations

Group	Benchmark	Median time	vs main
Grouping	GroupRules100	223521.907 ns/op	-3.8% (better)
Parsing	ParseGoOutput40	271256.522 ns/op	+2.5% (worse)
Validation	ValidateArtifact40	292918.230 ns/op	-1.9% (better)

Compared only with the latest matching platform in the main series.

[fennoai]

Review: v1.0.5 release bump

Clean, well-scoped patch release. The version field is bumped from 1.0.4 to 1.0.5 consistently across both package.json and the two synchronized locations in package-lock.json; no stale 1.0.4 reference remains.

Verified across all four review dimensions (quality, performance, security, documentation):

• Consistency — all three project-level version fields updated together.
• Dependencies — the lockfile dependency graph is untouched: re2js@2.8.6 and yaml@2.9.0 stay at exact pins, and all resolved URLs and integrity hashes are unchanged (no supply-chain impact).
• Docs — README pins the action via the floating major tag @v1, so a patch bump requires no README edits. No CHANGELOG exists in the repo historically, so its absence here is consistent with prior releases, not a regression.
• Performance — no runtime code paths affected.

No issues found. LGTM.