@sema-agent/core 2.0.1 → 2.2.0

package/dist/tools/web.js

@@ -577,6 +577,13 @@ export function createWebFetchSummarizer(brain, model) {
 const DEFAULT_SEARCH_TIMEOUT_MS = 30_000;
 const SEARCH_ERROR_EXCERPT_CHARS = 2048;
 const SEARCH_QUERY_MAX_CHARS = 2_000;
+const DETAILS_QUERY_ECHO_MAX_CHARS = 500;
+function clipQueryEcho(query) {
+    const cps = [...query];
+    return cps.length > DETAILS_QUERY_ECHO_MAX_CHARS
+        ? cps.slice(0, DETAILS_QUERY_ECHO_MAX_CHARS).join("") + `…[${cps.length} chars total]`
+        : query;
+}
 function classifySearchFailure(message) {
     const m = message.toLowerCase();
     const status = /(?:\bhttp\b|\bstatus\b|\bcode\b|\berror\b)\D{0,12}?(\d{3})\b/.exec(m)?.[1];
@@ -619,6 +626,18 @@ function hostMatchesDomain(host, domain) {
         return false;
     return h === d || h.endsWith("." + d);
 }
+const URL_SCHEME_PREFIX = /^([a-z][a-z0-9+.-]*):/i;
+function normalizeResultUrl(raw) {
+    const s = raw.trim();
+    if (!s)
+        return { url: raw, schemeAssumed: false };
+    if (s.startsWith("//"))
+        return { url: "https:" + s, schemeAssumed: true };
+    const scheme = URL_SCHEME_PREFIX.exec(s)?.[1];
+    if (scheme !== undefined && !scheme.includes("."))
+        return { url: s, schemeAssumed: false };
+    return { url: "https://" + s, schemeAssumed: true };
+}
 function webSearchResultAllowed(url, allowed, blocked) {
     let host;
     try {
@@ -662,7 +681,7 @@ export function createWebSearchTool(config) {
             const startedAt = Date.now();
             const failCard = (extra = {}) => ({
                 type: "web-search",
-                query,
+                query: clipQueryEcho(query),
                 results: [],
                 durationSeconds: (Date.now() - startedAt) / 1000,
                 searchCount: 0,
@@ -723,13 +742,18 @@ export function createWebSearchTool(config) {
                 clearTimeout(timer);
                 ctx.signal?.removeEventListener("abort", onOuterAbort);
             }
-            const usable = results.filter((r) => webSearchResultAllowed(r.url));
-            const droppedUnusable = results.length - usable.length;
+            const normalized = results.map((r) => {
+                const n = normalizeResultUrl(r.url);
+                return { ...r, url: n.url, schemeAssumed: n.schemeAssumed };
+            });
+            const usable = normalized.filter((r) => webSearchResultAllowed(r.url));
+            const droppedUnusable = normalized.length - usable.length;
             const domainFiltered = usable.filter((r) => webSearchResultAllowed(r.url, allowed_domains, blocked_domains));
             const droppedByDomain = usable.length - domainFiltered.length;
-            results = domainFiltered.slice(0, max);
+            const kept = domainFiltered.slice(0, max);
+            const schemeAssumed = kept.filter((r) => r.schemeAssumed).length;
             let clipped = false;
-            const shown = results.map((r) => {
+            const shown = kept.map((r) => {
                 const title = clipCodePoints(r.title, SEARCH_TITLE_MAX_CHARS);
                 const url = clipCodePoints(r.url, MAX_URL_LENGTH);
                 const snippet = clipCodePoints(r.snippet, SEARCH_SNIPPET_MAX_CHARS);
@@ -750,7 +774,13 @@ export function createWebSearchTool(config) {
             const clipNote = clipped
                 ? "\n\n[WebSearch: one or more result fields exceeded the display limits and were truncated — treat the text above as an excerpt, not the backend's full result.]"
                 : "";
-            const modelText = `${fenced}${dropNote}${domainNote}${clipNote}\n\nREMINDER: You MUST include the sources above in your response to the user using markdown hyperlinks.`;
+            const schemeNote = schemeAssumed > 0
+                ? `\n\n[WebSearch: ${schemeAssumed} result(s) came back with no URL scheme (a bare domain or a protocol-relative URL) and are shown normalized to https:// — that scheme is this tool's assumption, not the backend's claim.]`
+                : "";
+            const tailReminder = shown.length > 0
+                ? "REMINDER: You MUST include the sources above in your response to the user using markdown hyperlinks."
+                : "REMINDER: this search returned ZERO usable sources — there is nothing above to cite. Do not fabricate sources, URLs or citations; tell the user the search returned no usable results (the notes above say why, when there is a why) and answer from what you already know, or try a different query.";
+            const modelText = `${fenced}${dropNote}${domainNote}${schemeNote}${clipNote}\n\n${tailReminder}`;
             return {
                 content: modelText,
                 details: {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sema-agent/core",
-  "version": "2.0.1",
+  "version": "2.2.0",
   "description": "Stateless, task-oriented AI agent core",
   "type": "module",
   "license": "BUSL-1.1",
@@ -56,7 +56,8 @@
     "smoke": "tsx src/examples/smoke.ts",
     "eval": "tsx eval/run.ts",
     "prepare": "npm run build",
-    "prepublishOnly": "npm run build && npm run test"
+    "prepublishOnly": "npm run build && npm run test",
+    "gate:blackbox": "node scripts/run-blackbox-gate.mjs"
   },
   "dependencies": {
     "@modelcontextprotocol/sdk": "1.29.0",