CoderCup — initial public release

The continuous public benchmark for AI coding agents, refereed by TestSprite.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: yunhao-j

.github/ISSUE_TEMPLATE/bug.md

@@ -0,0 +1,22 @@
+---
+name: 🐛 Bug report
+about: Something on codercup.ai broken or wrong
+title: 'Bug: '
+labels: bug
+---
+
+**Where**
+
+<!-- Page URL or component -->
+
+**What happens**
+
+<!-- One paragraph -->
+
+**What should happen**
+
+<!-- One paragraph -->
+
+**Repro**
+
+<!-- Steps; browser + viewport if relevant -->

.github/ISSUE_TEMPLATE/new-driver.md

@@ -0,0 +1,36 @@
+---
+name: 🤖 New agent driver
+about: Propose adding a new coding agent to CoderCup
+title: 'New driver: <agent-name>'
+labels: new-driver
+---
+
+<!--
+  CoderCup is open to any AI coding agent that runs on a sandboxed
+  Linux host through a CLI. Onboarding is a ~30-line driver entry
+  in runners/drivers/<agent-slug>/ + contract review.
+
+  See runners/README.md#adding-a-new-driver for the full process.
+-->
+
+**Agent name + CLI invocation**
+
+<!-- e.g., "Aider — `aider --message-file <prompt>`" -->
+
+**Authentication model**
+
+<!-- Subscription / API key / OAuth / other -->
+
+**Public pricing**
+
+<!-- Link to the model's public rate card so we can populate
+     scoring/rates.ts -->
+
+**Streaming behavior**
+
+<!-- Does the CLI emit JSONL events to stdout? structured logs?
+     just plain text? -->
+
+**Maintainer**
+
+<!-- Who will be on call if this driver breaks? -->

.github/ISSUE_TEMPLATE/task-suggestion.md

@@ -0,0 +1,43 @@
+---
+name: 📋 Task suggestion
+about: Suggest a new requirement or surface to add to a CoderCup task spec
+title: 'Task suggestion: '
+labels: task-suggestion
+---
+
+<!--
+  The current task is World Cup 2026 (v1 → v2 in progress). New
+  requirements get folded into the next spec version + scored on the
+  next agent cohort run.
+
+  See task-spec/ for the current spec versions and tests/world-cup-v1/
+  for the registered TestSprite plans.
+-->
+
+**Surface to add**
+
+<!-- One sentence: what should the agent's deployed app expose? -->
+
+**Why this matters**
+
+<!-- One paragraph: what user need does it address, or what
+     engineering capability does it differentiate? -->
+
+**Suggested test plan (optional)**
+
+<!-- If you can sketch a TestSprite plan in the world-cup-v1 format:
+  {
+    "type": "frontend",
+    "name": "...",
+    "description": "...",
+    "priority": "p2",
+    "planSteps": [
+      { "type": "action", "description": "..." },
+      { "type": "assertion", "description": "..." }
+    ]
+  }
+-->
+
+**Acceptance**
+
+<!-- Concrete: when does this requirement count as met? -->

.github/workflows/ci.yml

@@ -0,0 +1,49 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+jobs:
+  test:
+    name: Unit tests + Next.js build
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-node@v4
+        with:
+          node-version: '20.10.0'
+          cache: 'npm'
+
+      - name: Install dependencies
+        run: npm ci
+
+      - name: Type check + run unit tests
+        run: npm test -- --run
+
+      - name: Build the leaderboard (Next.js static export)
+        run: npm run build
+
+  # CDK tsc-only check (no synth — RunnerStack does a VPC lookup that needs
+  # AWS credentials, and we don't want to gate CI on OIDC setup right now).
+  cdk-typecheck:
+    name: CDK type check (infra)
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: infra/cdk
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-node@v4
+        with:
+          node-version: '20.10.0'
+
+      - name: Install CDK deps
+        run: npm ci
+
+      - name: tsc --noEmit
+        run: npx tsc --noEmit

.gitignore

@@ -0,0 +1,62 @@
+# dependencies
+/node_modules
+/.pnp
+.pnp.*
+.yarn/*
+!.yarn/patches
+!.yarn/plugins
+!.yarn/releases
+!.yarn/versions
+
+# testing
+/coverage
+
+# next.js
+/.next/
+/out/
+
+# production
+/build
+
+# misc
+.DS_Store
+*.pem
+
+# debug
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+.pnpm-debug.log*
+
+# env files (DO NOT COMMIT SECRETS)
+.env
+.env*.local
+.env.development
+.env.production
+
+# typescript
+*.tsbuildinfo
+next-env.d.ts
+
+# editor / OS
+.idea
+.vscode
+.cursor
+*.swp
+
+# AWS CDK
+infra/cdk.out
+infra/cdk.context.json
+
+# runner artifacts (local-only; production drops to S3)
+/runs/
+
+# Local Claude scheduler state
+.claude/
+
+# TestSprite local run artifacts (videos/snapshots — large, not source)
+.testsprite/
+err.txt
+
+# internal ops docs + orchestration harness (not part of the open-source release)
+docs-internal/