@agentskillshub/cli 0.1.0 → 0.2.1

package/bin/ash.mjs

@@ -42,6 +42,9 @@ const GRADE = {
   unknown: { label: "UNAUDITED", mark: "⚪" },
 };
 
+// CJK detection — Chinese queries have no word boundaries, so we bigram them.
+const CJK = /[一-鿿]/;
+
 // ─── tiny ANSI (auto-off when not a TTY / NO_COLOR) ──────────────────────────
 const tty = process.stdout.isTTY && !process.env.NO_COLOR;
 const c = (code, s) => (tty ? `\x1b[${code}m${s}\x1b[0m` : s);
@@ -143,13 +146,26 @@ function scoreRow(row, tokens) {
   const full = (row.f || "").toLowerCase();
   const desc = (row.d || "").toLowerCase();
   const tags = (row.t || []).join(" ").toLowerCase();
+  // Bilingual scenario keywords (field `w`) — lets Chinese queries match
+  // English-only repos via our curated scenario titles, and ranks
+  // scenario-relevant skills higher. Empty/undefined on older indexes.
+  const scen = (row.w || "").toLowerCase();
   let score = 0;
   for (const tok of tokens) {
     if (name === tok) score += 50;
     else if (name.includes(tok)) score += 20;
     if (full.includes(tok)) score += 8;
+    if (scen.includes(tok)) score += 12;
     if (tags.includes(tok)) score += 10;
     if (desc.includes(tok)) score += 5;
+    // CJK compound queries arrive as one space-less token ("抓取网站").
+    // Fall back to 2-char windows so a keyword like "抓取" still matches.
+    if (CJK.test(tok) && tok.length >= 3) {
+      for (let i = 0; i + 2 <= tok.length; i++) {
+        const bg = tok.slice(i, i + 2);
+        if (scen.includes(bg)) { score += 9; break; }
+      }
+    }
   }
   if (score === 0) return -1; // matched nothing
   return score + (row.q || 0) / 20 + Math.min(row.s, 50000) / 25000; // quality + popularity tiebreak
@@ -165,9 +181,22 @@ function applyFilters(skills, f) {
   });
 }
 
+/** Tokenize a query. Splits on whitespace AND at latin↔CJK boundaries, so a
+ *  glued mixed query like "ppt制作" becomes ["ppt", "制作"] (otherwise it's one
+ *  token that matches nothing). Pure-CJK compounds still rely on the bigram
+ *  fallback in scoreRow. */
+function tokenize(q) {
+  return q
+    .toLowerCase()
+    .replace(/([a-z0-9])([一-鿿])/g, "$1 $2")
+    .replace(/([一-鿿])([a-z0-9])/g, "$1 $2")
+    .split(/\s+/)
+    .filter(Boolean);
+}
+
 function runSearch(index, args) {
   const f = parseFilters(args);
-  const tokens = f.query.toLowerCase().split(/\s+/).filter(Boolean);
+  const tokens = tokenize(f.query);
   const pool = applyFilters(index.skills, f);
   const ranked = pool
     .map((r) => ({ r, score: scoreRow(r, tokens) }))
@@ -247,7 +276,7 @@ function runAudit(index, args) {
     return console.log(JSON.stringify({ ...expand(row), in_catalog: true, verdict: VERDICT[row.g] || VERDICT.unknown, tier_note: "Basic (free). 5-dimension deep audit + any GitHub URL → Pro." }, null, 2));
   }
   console.log(`\n${bold(cyan(row.f))}  ${yellow(starStr(row.s))}${row.o ? green("  ✓ official") : ""}`);
-  console.log(`Security: ${gradeBadge(row.g)}     Quality: ${row.q}/100`);
+  console.log(`Security: ${gradeBadge(row.g)}     Quality: ${Math.round(row.q ?? 0)}/100`);
   if (row.d) console.log(dim(`\n  ${row.d}`));
   console.log(`\n${bold("Basic verdict")} ${dim("(free tier)")}`);
   console.log(`  ${VERDICT[row.g] || VERDICT.unknown}`);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@agentskillshub/cli",
-  "version": "0.1.0",
+  "version": "0.2.1",
   "description": "Search, audit, and install open-source AI agent skills & MCP servers from the terminal — security-graded, quality-scored.",
   "type": "module",
   "bin": {