feat: curated public benchmark dataset and leaderboard

.github/workflows/validate.yml

@@ -53,3 +53,22 @@ jobs:
 
       - name: Validate eval schemas
         run: bun apps/cli/dist/cli.js validate 'examples/features/**/evals/**/*.eval.yaml' 'examples/features/**/*.EVAL.yaml'
+
+  benchmark-results:
+    name: Validate Benchmark Results
+    runs-on: ubuntu-latest
+    if: >-
+      contains(github.event.pull_request.title, 'benchmark') ||
+      contains(join(github.event.pull_request.labels.*.name, ','), 'benchmark') ||
+      github.event_name == 'push'
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/setup-bun
+
+      - name: Validate SWE-bench Lite result JSON files
+        run: |
+          if ls benchmarks/swe-bench-lite/results/*.json 1> /dev/null 2>&1; then
+            bun benchmarks/swe-bench-lite/validate-result.ts benchmarks/swe-bench-lite/results/*.json
+          else
+            echo "No result files found — skipping"
+          fi

apps/web/src/components/Lander.astro

@@ -14,6 +14,7 @@
     </a>
     <div class="av-nav-links">
       <a href="/docs/">Docs</a>
+      <a href="/leaderboard">Leaderboard</a>
       <a href="https://github.com/EntityProcess/agentv" target="_blank" rel="noopener noreferrer">GitHub</a>
       <button class="av-nav-pill" data-command="npm install -g agentv">
         <code>npm install -g agentv</code>
@@ -118,6 +119,19 @@
     </div>
   </section>
 
+  <!-- Leaderboard CTA Section -->
+  <section class="av-features" style="border-top: 1px solid rgba(255,255,255,0.04);">
+    <div class="av-container" style="text-align:center;">
+      <h2 class="av-section-heading">Public Leaderboard</h2>
+      <p style="color:#94a3b8; max-width:560px; margin:0 auto 1.5rem; font-size:0.95rem;">
+        SWE-bench Lite results with richer metrics — cost efficiency, tool usage, and Pareto-optimal rankings. See how models actually compare.
+      </p>
+      <a href="/leaderboard" class="av-btn-primary" style="display:inline-block; padding:0.75rem 2rem; font-size:0.9rem;">
+        View Leaderboard →
+      </a>
+    </div>
+  </section>
+
   <!-- Quick Start Section -->
   <section class="av-quickstart">
     <div class="av-container">