Merge remote-tracking branch 'origin/main' into claude/github-issue-11-9hCO5

# Conflicts:
#	src/server/preflight.ts

Author: slaFFik

.github/FUNDING.yml

@@ -0,0 +1,15 @@
+# These are supported funding model platforms
+
+github: [slaFFik]
+patreon: # Replace with a single Patreon username
+open_collective: # Replace with a single Open Collective username
+ko_fi: # Replace with a single Ko-fi username
+tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
+community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
+liberapay: # Replace with a single Liberapay username
+issuehunt: # Replace with a single IssueHunt username
+lfx_crowdfunding: # Replace with a single LFX Crowdfunding project-name e.g., cloud-foundry
+polar: # Replace with a single Polar username
+buy_me_a_coffee: # Replace with a single Buy Me a Coffee username
+thanks_dev: # Replace with a single thanks.dev username
+custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']

.github/workflows/ci.yml

@@ -50,59 +50,44 @@ jobs:
       - uses: actions/checkout@v6
       - uses: ./.github/actions/setup
         with:
-          # ESLint v10 requires ^20.19.0 || ^22.13.0 || >=24, so pick a Node
-          # that satisfies it (the typecheck/build/test matrix stays on the
-          # project's engines floor 20.19/22.12 to test the actual runtime).
           node-version: '22.13.x'
       - run: npm run lint
       - run: npm run format:check
 
   typecheck:
-    name: Typecheck (Node ${{ matrix.node-version }})
+    name: Typecheck
     runs-on: ubuntu-latest
     timeout-minutes: 10
     if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    strategy:
-      fail-fast: false
-      matrix:
-        node-version: ['20.19.x', '22.12.x']
     steps:
       - uses: actions/checkout@v6
       - uses: ./.github/actions/setup
         with:
-          node-version: ${{ matrix.node-version }}
+          node-version: '22.13.x'
       - run: npm run typecheck
 
   build:
-    name: Build (Node ${{ matrix.node-version }})
+    name: Build
     runs-on: ubuntu-latest
     timeout-minutes: 10
     if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    strategy:
-      fail-fast: false
-      matrix:
-        node-version: ['20.19.x', '22.12.x']
     steps:
       - uses: actions/checkout@v6
       - uses: ./.github/actions/setup
         with:
-          node-version: ${{ matrix.node-version }}
+          node-version: '22.13.x'
       - run: npm run build
 
   test:
-    name: Test (Node ${{ matrix.node-version }})
+    name: Test
     runs-on: ubuntu-latest
     timeout-minutes: 15
     if: github.event_name != 'pull_request' || github.event.pull_request.draft == false
-    strategy:
-      fail-fast: false
-      matrix:
-        node-version: ['20.19.x', '22.12.x']
     steps:
       - uses: actions/checkout@v6
       - uses: ./.github/actions/setup
         with:
-          node-version: ${{ matrix.node-version }}
+          node-version: '22.13.x'
       - run: npm run test:preflight
       - run: npm run test:judge
       - run: npm run test:runner

.nvmrc

@@ -0,0 +1 @@
+22

package-lock.json

@@ -29,7 +29,7 @@
     "test:sandbox": "tsx test/sandbox.spec.ts"
   },
   "engines": {
-    "node": "^20.19.0 || >=22.12.0"
+    "node": ">=22.13.0"
   },
   "dependencies": {
     "@radix-ui/react-tooltip": "^1.2.8",
@@ -42,7 +42,7 @@
   },
   "devDependencies": {
     "@eslint/js": "^10.0.1",
-    "@types/node": "^20.19.39",
+    "@types/node": "^22.19.17",
     "@types/proper-lockfile": "^4.1.4",
     "@types/react": "^19.2.14",
     "@types/react-dom": "^19.2.3",

package.json

@@ -3,7 +3,7 @@ import { promisify } from 'node:util';
 import { homedir, tmpdir } from 'node:os';
 import { join } from 'node:path';
 import { pathExists, atomicWriteFile, ensureDir, readJsonIfExists } from './fsUtil.js';
-import { PROJECT_MARKERS, STORAGE_DIR_NAME } from '@shared/constants.js';
+import { JUDGE_MODEL, PROJECT_MARKERS, STORAGE_DIR_NAME } from '@shared/constants.js';
 import { readdir, readFile, stat, unlink, writeFile } from 'node:fs/promises';
 import lockfile from 'proper-lockfile';
 import { log } from './log.js';
@@ -114,7 +114,7 @@ export async function authSmokeTest(bin: string): Promise<void> {
     '-p',
     'ping',
     '--model',
-    'haiku',
+    JUDGE_MODEL,
     '--output-format',
     'json',
     '--tools',

src/server/preflight.ts

@@ -382,6 +382,10 @@ export class Runner extends EventEmitter {
       },
     );
 
+    if (typeof raw.model === 'string' && raw.model.length > 0) {
+      this.input.initialConfig.resolvedModel = raw.model;
+    }
+
     const memoryPaths = raw.memory_paths as Record<string, unknown> | undefined;
     const auto = memoryPaths?.auto;
     if (typeof auto !== 'string' || auto.length === 0) return;

src/server/runner.ts

@@ -3,6 +3,7 @@ import { spawn } from 'node:child_process';
 import { pathExists } from './fsUtil.js';
 import { join } from 'node:path';
 import { readdir } from 'node:fs/promises';
+import { SLUG_MODEL } from '@shared/constants.js';
 import { log } from './log.js';
 
 const MAX_SLUG_LENGTH = 32;
@@ -75,9 +76,9 @@ export async function deriveSlug(
   }
 
   if (!slugBase) {
-    const haikuSlug = await tryHaikuSlug(input.variantContent, input.claudeBin);
-    if (haikuSlug && isSafeSlugFragment(haikuSlug)) {
-      slugBase = haikuSlug;
+    const modelSlug = await tryModelSlug(input.variantContent, input.claudeBin);
+    if (modelSlug && isSafeSlugFragment(modelSlug)) {
+      slugBase = modelSlug;
       source = 'haiku';
     }
   }
@@ -110,7 +111,7 @@ export function slugStoragePath(storageRoot: string, folderName: string): string
   return join(storageRoot, folderName);
 }
 
-async function tryHaikuSlug(content: string, claudeBin: string): Promise<string | null> {
+async function tryModelSlug(content: string, claudeBin: string): Promise<string | null> {
   const prompt =
     'Produce a 2-4 word kebab-case slug summarizing this variant. Output only the slug, no quotes, no explanation. Example: "concise-style" or "verbose-debugging". Variant content follows:\n\n' +
     content.slice(0, 4_000);
@@ -129,7 +130,7 @@ async function tryHaikuSlug(content: string, claudeBin: string): Promise<string
         '-p',
         prompt,
         '--model',
-        'haiku',
+        SLUG_MODEL,
         '--output-format',
         'json',
         '--tools',
@@ -147,7 +148,7 @@ async function tryHaikuSlug(content: string, claudeBin: string): Promise<string
     let buf = '';
     proc.stdout.on('data', (d) => (buf += d.toString()));
     proc.stderr.on('data', (d) => {
-      log.debug('haiku-slug stderr', { text: d.toString().slice(0, 200) });
+      log.debug('model-slug stderr', { text: d.toString().slice(0, 200) });
     });
 
     const timer = setTimeout(() => {
@@ -157,7 +158,7 @@ async function tryHaikuSlug(content: string, claudeBin: string): Promise<string
 
     proc.on('error', (err) => {
       clearTimeout(timer);
-      log.warn('haiku-slug spawn error', { error: err.message });
+      log.warn('model-slug spawn error', { error: err.message });
       finish(null);
     });
 

src/server/slug.ts

@@ -20,7 +20,17 @@ export const PROJECT_MARKERS = [
   'pyproject.toml',
 ];
 
-export const JUDGE_MODEL = 'haiku';
+// Pin to a concrete model ID, not the `haiku` CLI alias: aliases can be
+// repointed to a future generation or removed, which would silently shift
+// score baselines or break the judge entirely (issue #13). Bump manually
+// when a new Haiku ships and rebaselining is acceptable.
+export const JUDGE_MODEL = 'claude-haiku-4-5';
+
+// Model used to auto-generate run-folder slugs from variant content. Kept
+// separate from JUDGE_MODEL so the slug generator can move independently —
+// slugs are non-critical (they fall back gracefully) and don't anchor any
+// historical comparison the way judge scores do.
+export const SLUG_MODEL = 'claude-haiku-4-5';
 
 // Judge input budgets (plan § Judge flow step 2).
 export const JUDGE_PROMPT_CAP_BYTES = 4 * 1024;

src/shared/constants.ts

@@ -26,6 +26,11 @@ export const RunConfigSchema = z.object({
   promptSha256: z.string(),
   prompt: z.string(),
   model: ModelIdSchema,
+  // Concrete model identifier the CLI reports in its `system_init` event.
+  // `model` above is what the user selected (often an alias like `haiku`);
+  // this captures what actually ran, so reruns across an alias bump remain
+  // comparable. Optional/nullable for back-compat with pre-existing configs.
+  resolvedModel: ModelIdSchema.nullable().optional(),
   mode: ModeSchema,
   status: RunStatusSchema,
   startedAt: z.string(),