diff --git a/README.md b/README.md
index 035ce359..726417ab 100644
--- a/README.md
+++ b/README.md
@@ -69,8 +69,13 @@ The same CLI also covers the work before the code — requirements discovery, ta
 decomposition, and delegating each task to **Claude Code, Codex, OpenCode,
 Kilocode** or its built-in agent. See [Features](#features).
 
-**Proof of substance.** 5,000+ tests with an 80% coverage floor enforced in CI,
-and a [sample evidence report](https://codeframe.sh/sample-evidence-report.html)
+<!-- Figures rounded DOWN so they stay true as the suite grows (CI on main @ 507c6e0c, 2026-10-28: 6,035 passed, 86.13% coverage, fail_under = 80).
+     Coverage is stated as the floor CI enforces, never the measured value:
+     tests/test_lifecycle_gates_948.py holds README.md to fail_under (#948).
+     Keep in step with the stats block in site/index.html. -->
+**Proof of substance.** 6,000+ automated tests passing in CI, an 80% test coverage
+floor CI enforces, and a
+[sample evidence report](https://codeframe.sh/sample-evidence-report.html)
 showing what a run actually produces.
 
 > [!NOTE]
@@ -410,7 +415,7 @@ uv run pytest tests/ --ignore=tests/e2e -m "not lifecycle"   # the CI gate
 uv run ruff check .
 ```
 
-The suite has 5,000+ tests, and contributions need 80% test coverage, the floor CI enforces. Code
+The suite has 6,000+ tests, and contributions need 80% test coverage, the floor CI enforces. Code
 standards are ruff, type hints and tests with every change. During the beta,
 feature ideas go to [Discussions → Ideas](https://github.com/frankbria/codeframe/discussions/categories/ideas)
 before code. See [CONTRIBUTING.md](CONTRIBUTING.md) for what is stable and how
diff --git a/site/index.html b/site/index.html
index 000cafa4..ec0fa790 100644
--- a/site/index.html
+++ b/site/index.html
@@ -90,6 +90,12 @@ <h1>Your coding agent can&rsquo;t quietly undo what you already decided.</h1>
       <figcaption>A real session on codeframe 0.9.7. Asked for a change that would undo a past
         decision, Claude finds the locked requirement and keeps it.</figcaption>
     </figure>
+
+    <!-- COPY: answers the first objection an engineer has (README first screen). -->
+    <p class="tdd"><strong>TDD protects your code. This protects your tests.</strong> Your agent
+      can edit a test as easily as it can edit the code, and when a test is inconvenient that is
+      often what it edits. A requirement is a test it cannot quietly drop: releasing one takes a
+      waiver with a name, a reason and an expiry date.</p>
   </header>
 
   <section aria-labelledby="who">
@@ -102,10 +108,13 @@ <h2 id="who">Who built it</h2>
   <section aria-labelledby="substance">
     <h2 id="substance">Proof of substance</h2>
     <!-- COPY: figures. Marketing-rounded DOWN so they stay true as the suite grows
-         (2026-10-14, d26a5c8: 7,891 backend tests, 87.42% coverage, fail_under = 80). -->
+         (CI on main @ 507c6e0c, 2026-10-28: 6,035 passed, 86.13% coverage, fail_under = 80).
+         Coverage is stated as the floor CI enforces, never the measured value:
+     tests/test_lifecycle_gates_948.py holds README.md to fail_under (#948).
+         Keep in step with "Proof of substance" in README.md. -->
     <div class="stats">
-      <div class="stat"><b>7,500+</b><span>automated tests passing in CI</span></div>
-      <div class="stat"><b>85%+</b><span>test coverage, with an 80% floor CI enforces</span></div>
+      <div class="stat"><b>6,000+</b><span>automated tests passing in CI</span></div>
+      <div class="stat"><b>80%</b><span>test-coverage floor, enforced in CI</span></div>
     </div>
     <p>Every check leaves evidence. See a <a href="sample-evidence-report.html">sample evidence report</a>
       from the session above, made with <code>cf proof export</code>.</p>