Proposal: </h1>
<blockquote>
<p>Enhancement proposals (KEP/RFC-style) are how larger changes and milestone plans are
proposed. Move this file from <code>draft/</code> to <code>accepted/</code> when accepted; an accepted proposal
is never edited — a later proposal supersedes it, recorded in <code>../proposals/README.md</code>.
If a decision is foundational rather than a change, it may instead warrant an ADR.</p>
</blockquote>
<h2 id="motivation">Motivation</h2>
<p>What problem does this solve? Why now? What does the user see?</p>
<h2 id="design">Design</h2>
<p>The proposed change. Be concrete about the affected components and about any effect on the
properties the threat model says must hold — those carry the highest bar.</p>
<h2 id="alternatives-considered">Alternatives considered</h2>
<p>What else was weighed, and why this over those?</p>
<h2 id="graduation-criteria">Graduation criteria</h2>
<p>How do we know it is done and correct? Which tests, which quality scenarios (architecture
§10), and what definition of done?</p>
<h2 id="security-impact">Security impact</h2>
<p>Which untrusted inputs from <code>.oss-scanner/threat_model.md</code> does this touch? Does it add a
new one, or change what is trusted? Update the threat model in the same change if so.</p>
<h2 id="backward-compatibility">Backward compatibility</h2>
<p>Effect on existing data, configs, deployments and the public interface.</p>
<h2 id="open-questions">Open questions</h2>
<p>Mark unresolved points explicitly.</p>
</article>
</main>
</div>
<div class="wrap">
<footer>
<div class="foot-row">
<nav class="foot-links" aria-label="footer">
<a href="https://github.com/eduralph/pdca-harness">github</a>
<a href="/guide/">guide</a>
<a href="/spec/">spec</a>
<a href="/design/">design</a>
<a href="/governance/">governance</a>
</nav>
<div class="foot-meta">Apache-2.0 · Agentic AI · model agnostic</div>
</div>
</footer>
</div>
<!-- SCRIPTS: render_site.py injects the Mermaid include only on pages that
contain a diagram; empty otherwise. -->
</body>
</html>