Merge pull request #372 from ShiroKSH/fix/eval-validation-windows

fix: harden eval and command validation

.gitattributes

@@ -0,0 +1 @@
+* text=auto eol=lf

README.md

@@ -339,7 +339,7 @@ agent-skills/
 │   ├── shipping-and-launch/           #   Ship
 │   └── using-agent-skills/            #   Meta: how to use this pack
 ├── agents/                            # 4 specialist personas
-├── references/                        # 5 supplementary checklists
+├── references/                        # 7 supplementary checklists
 ├── hooks/                             # Session lifecycle hooks
 ├── .claude/commands/                  # 8 slash commands (Claude Code)
 ├── .gemini/commands/                  # 8 slash commands (Gemini CLI)

docs/copilot-setup.md

@@ -7,7 +7,7 @@
 Copilot supports creating agent skills using a `.github/skills`, `.claude/skills`, or `.agents/skills` directory in your repository.
 
 ```bash
-mkdir -p .github
+mkdir -p .github/skills/test-driven-development .github/skills/code-review-and-quality
 
 # Create files for essential skills
 cat /path/to/agent-skills/skills/test-driven-development/SKILL.md > .github/skills/test-driven-development/SKILL.md

docs/getting-started.md

@@ -132,6 +132,9 @@ The `references/` directory contains supplementary checklists:
 | `performance-checklist.md` | performance-optimization |
 | `security-checklist.md` | security-and-hardening |
 | `accessibility-checklist.md` | frontend-ui-engineering |
+| `definition-of-done.md` | all skills / every change |
+| `observability-checklist.md` | observability-and-instrumentation |
+| `orchestration-patterns.md` | context-engineering |
 
 Load a reference when you need detailed patterns beyond what the skill covers.
 

scripts/run-evals.js

@@ -154,7 +154,7 @@ function loadSkills() {
     const file = path.join(SKILLS_DIR, dir, 'SKILL.md');
     if (!fs.existsSync(file)) continue;
     const src = fs.readFileSync(file, 'utf8');
-    const m = src.match(/^---\n([\s\S]*?)\n---/);
+    const m = src.match(/^---[ \t]*\r?\n([\s\S]*?)\r?\n---[ \t]*(?:\r?\n|$)/);
     if (!m) continue;
     const name = (m[1].match(/^name:\s*(.+)$/m) || [])[1];
     const description = (m[1].match(/^description:\s*(.+)$/m) || [])[1];
@@ -178,6 +178,19 @@ function loadCases() {
     });
 }
 
+function resolveFixturePath(root, rel) {
+  if (path.isAbsolute(rel)) {
+    throw new Error(`fixture path must be relative: ${rel}`);
+  }
+  const resolvedRoot = path.resolve(root);
+  const resolvedPath = path.resolve(resolvedRoot, rel);
+  const back = path.relative(resolvedRoot, resolvedPath);
+  if (back === '' || back === '..' || back.startsWith(`..${path.sep}`) || path.isAbsolute(back)) {
+    throw new Error(`fixture path escapes workspace: ${rel}`);
+  }
+  return resolvedPath;
+}
+
 // ---------- tier 2 ----------
 
 function runDeterministic() {
@@ -331,11 +344,11 @@ function materializeWorkspace(ev) {
   // agent has real code to operate on rather than describing what it would do.
   const workspace = fs.mkdtempSync(path.join(os.tmpdir(), 'agent-skills-eval-'));
   for (const rel of ev.files || []) {
-    const src = path.join(FIXTURES_DIR, rel);
+    const src = resolveFixturePath(FIXTURES_DIR, rel);
     if (!fs.existsSync(src)) {
       throw new Error(`fixture listed in files[] not found: evals/fixtures/${rel}`);
     }
-    const dest = path.join(workspace, rel);
+    const dest = resolveFixturePath(workspace, rel);
     fs.mkdirSync(path.dirname(dest), { recursive: true });
     fs.cpSync(src, dest, { recursive: true });
   }

scripts/validate-commands.js

@@ -145,6 +145,21 @@ function main() {
     const descGemini = byTool.gemini[tomlStem];
     const descAgy    = byTool.antigravity[tomlStem];
 
+    const malformed = [
+      ['.claude/commands', byTool.claude, claudeStem],
+      ['.gemini/commands', byTool.gemini, tomlStem],
+      ['commands/', byTool.antigravity, tomlStem],
+    ].filter(([, commands, stem]) => Object.prototype.hasOwnProperty.call(commands, stem) && commands[stem] == null);
+
+    if (malformed.length) {
+      console.log(`  ✗  ${claudeStem}`);
+      for (const [toolDir, , stem] of malformed) {
+        console.log(`       ${toolDir}/${stem} — missing or malformed description`);
+      }
+      errors++;
+      continue;
+    }
+
     if (descClaude == null || descGemini == null || descAgy == null) {
       // Missing file already flagged by parity check
       continue;