[sentomk]

Summary

Adds a project skill that describes the release workflow used for v0.9.4 and v0.9.5, so a release can be prepared and published from one instruction. The skill is documentation for the agent; it changes no library code.

Changes

• .claude/skills/release/SKILL.md: release PR steps (changelog, performance note with a measurement method, version bumps, registration), post-merge steps (tag, GitHub release, portfile SHA512), and the sentomk/vcpkg and microsoft/vcpkg update.
• States the standing rules: single-commit PRs, Conventional Commits, Summary / Changes / Testing PR bodies, no attribution trailers.

Testing

• Docs only. The steps were checked against the v0.9.4 release PR and the merged microsoft/vcpkg PRs for patternia.
• Commit message checked with scripts/check_commit_msg.py.

[biobo-bot]

Thanks, this is a clear and well-grounded write-up of the release process. I spot-checked the paths and conventions it references against main (scripts/check_commit_msg.py, the **Summary** / **Changes** / **Testing** check in .github/workflows/validate.yml, packaging/vcpkg/*, docs/changelog/releases.md, docs/performance/index.md, mkdocs.yml, the PTN_BUILD_TESTS option and the README GIT_TAG v0.9.5), and they all line up. No library code is touched, so there's nothing to break at build time.

Approving. I left one small non-blocking note on the vcpkg step ordering in Phase 2, step 4, which reads a little out of sequence as written.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[github-actions]

Benchmark comparison (informational)

Base 6ba27eb vs PR 204ccac, run one after the other on the same GitHub-hosted runner. This does not block merging.

Noise estimate: the 20 control rows below do not run Patternia code, so they should not change between base and PR. As a group they moved by a median of -0.6% (run order and machine drift); that offset is divided out of the Patternia deltas (Adjusted). The controls then scatter by a median of 2.0%. Adjusted differences under 5.0% are treated as noise.

Patternia: 8 benchmarks, 0 slower, 0 faster, 8 within noise.

Benchmark	Base (ns)	PR (ns)	Delta	Adjusted	Verdict
PatterniaPipe CommandParser	1.658	1.571	-5.2%	-4.6%	within noise
PatterniaPipe PacketMixed	1.489	1.525	+2.4%	+3.1%	within noise
PatterniaPipe PacketMixedHeavyBind	1.469	1.477	+0.5%	+1.2%	within noise
PatterniaPipe ProtocolRouter	1.343	1.288	-4.1%	-3.5%	within noise
PatterniaPipe VariantAltHot	1.056	1.051	-0.5%	+0.1%	within noise
PatterniaPipe VariantAltHotGuarded	1.131	1.097	-3.0%	-2.3%	within noise
PatterniaPipe VariantMixed	1.229	1.270	+3.3%	+4.0%	within noise
PatterniaPipe VariantMixedGuarded	1.095	1.097	+0.1%	+0.8%	within noise

Control rows (not Patternia code)

Benchmark	Base (ns)	PR (ns)	Delta
IfElse CommandParser	1.567	1.500	-4.3%
IfElse ProtocolRouter	1.941	1.932	-0.5%
Sequential VariantAltHot	1.177	1.252	+6.3%
Sequential VariantAltHotGuarded	1.263	1.242	-1.6%
Sequential VariantMixed	0.847	0.869	+2.6%
Sequential VariantMixedGuarded	1.045	1.052	+0.7%
StdVisit CommandParser	1.069	1.065	-0.4%
StdVisit ProtocolRouter	1.285	1.302	+1.3%
StdVisit VariantAltHot	1.072	1.045	-2.6%
StdVisit VariantAltHotGuarded	1.121	1.102	-1.6%
StdVisit VariantMixed	1.248	1.283	+2.7%
StdVisit VariantMixedGuarded	1.735	1.670	-3.7%
SwitchIndex VariantAltHot	1.187	1.177	-0.8%
SwitchIndex VariantAltHotGuarded	1.232	1.265	+2.6%
SwitchIndex VariantMixed	0.775	0.764	-1.4%
SwitchIndex VariantMixedGuarded	0.851	0.871	+2.4%
Switch CommandParser	1.511	1.471	-2.7%
Switch PacketMixed	1.087	1.073	-1.3%
Switch PacketMixedHeavyBind	1.353	1.411	+4.3%
Switch ProtocolRouter	1.319	1.287	-2.4%

Raw results and the comparison chart are attached to this workflow run as an artifact. Shared runners are noisy, so use this as a trend, not a measurement.