<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Zensu Blog</title><description>Field notes on AI-native product lifecycle management.</description><link>https://zensu.dev/</link><item><title>The release gate: features that ship when they&apos;re real, not when you say so</title><link>https://zensu.dev/blog/release-gate-earned-not-claimed/</link><guid isPermaLink="true">https://zensu.dev/blog/release-gate-earned-not-claimed/</guid><description>Status isn&apos;t a label you set — it&apos;s a gate you earn. Docs, coverage, journey health, and a risk-proportional security bar block the transition to released, so &apos;done&apos; means verified instead of claimed.</description><pubDate>Mon, 29 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Ghost-scan: turn an existing repo into a tracked feature graph</title><link>https://zensu.dev/blog/ghost-scan-existing-repo/</link><guid isPermaLink="true">https://zensu.dev/blog/ghost-scan-existing-repo/</guid><description>Adopt Zensu on a brownfield codebase without a rewrite — a multi-perspective scan surfaces undocumented features, tests, and docs as scored candidates, then apply auto-links them and drafts the user journeys that gate your first release.</description><pubDate>Wed, 03 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Five reviewers, one install: parallel multi-perspective code review</title><link>https://zensu.dev/blog/five-reviewers-one-install/</link><guid isPermaLink="true">https://zensu.dev/blog/five-reviewers-one-install/</guid><description>Conventions, bugs, architecture, tests, security — five read-only reviewers run in parallel on every change, then merge into a single consolidated report.</description><pubDate>Fri, 29 May 2026 00:00:00 GMT</pubDate></item><item><title>TDD you can&apos;t skip: a phase-gate inside your coding agent</title><link>https://zensu.dev/blog/tdd-you-cant-skip/</link><guid isPermaLink="true">https://zensu.dev/blog/tdd-you-cant-skip/</guid><description>Zensu&apos;s strict TDD flow is opt-in: turn it on and the harness blocks any edit without a failing test first. It costs about 30% more tokens — and on regression-heavy code, it&apos;s worth every one.</description><pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate></item></channel></rss>