diff --git a/packages/coding-agent/src/extensions/tool-search/tool.ts b/packages/coding-agent/src/extensions/tool-search/tool.ts
index 1841ebd97..b9fa4ade0 100644
--- a/packages/coding-agent/src/extensions/tool-search/tool.ts
+++ b/packages/coding-agent/src/extensions/tool-search/tool.ts
@@ -115,7 +115,28 @@ export function createToolSearchDocument(
 	return { name: tool.name, text: parts.filter((part) => part.trim()).join(" ") };
 }
 
-/** Okapi BM25 with the usual parameters. Ties keep document order. */
+/** Lowercase with characters invalid in script identifiers as `_`, so `dev-radius` matches `dev_radius`. */
+function normalizeName(name: string): string {
+	return name
+		.trim()
+		.toLowerCase()
+		.replace(/[^a-z0-9_$]/g, "_");
+}
+
+/**
+ * Whether `query` names the tool: its full name or a suffix after a `__` separator, so `thread_read`
+ * and `app__thread_read` both name `mcp__app__thread_read`.
+ */
+function isNameQuery(query: string, name: string): boolean {
+	const normalized = normalizeName(name);
+	return normalized === query || normalized.endsWith(`__${query}`);
+}
+
+/**
+ * Okapi BM25 with the usual parameters. A query that names a tool ranks that tool first, because
+ * length normalization can otherwise rank a short description sharing the name's words above it.
+ * Ties keep document order.
+ */
 export class Bm25Ranker implements ToolRanker {
 	private readonly k1: number;
 	private readonly b: number;
@@ -141,7 +162,8 @@ export class Bm25Ranker implements ToolRanker {
 				return [term, Math.log(1 + (documents.length - frequency + 0.5) / (frequency + 0.5))] as const;
 			}),
 		);
-		const matches: ToolSearchMatch[] = [];
+		const nameQuery = normalizeName(query);
+		const matches: (ToolSearchMatch & { named: boolean })[] = [];
 		documents.forEach((document, index) => {
 			let score = 0;
 			for (const term of queryTerms) {
@@ -150,9 +172,12 @@ export class Bm25Ranker implements ToolRanker {
 				const norm = this.k1 * (1 - this.b + (this.b * lengths[index]) / averageLength);
 				score += (idf.get(term) ?? 0) * ((count * (this.k1 + 1)) / (count + norm));
 			}
-			if (score > 0) matches.push({ name: document.name, score });
+			if (score > 0) matches.push({ name: document.name, score, named: isNameQuery(nameQuery, document.name) });
 		});
-		return matches.sort((a, b) => b.score - a.score).slice(0, limit);
+		return matches
+			.sort((a, b) => Number(b.named) - Number(a.named) || b.score - a.score)
+			.slice(0, limit)
+			.map(({ name, score }) => ({ name, score }));
 	}
 }
 
diff --git a/packages/coding-agent/test/tool-search.test.ts b/packages/coding-agent/test/tool-search.test.ts
index ba07448a0..fd55b5c56 100644
--- a/packages/coding-agent/test/tool-search.test.ts
+++ b/packages/coding-agent/test/tool-search.test.ts
@@ -53,6 +53,29 @@ describe("Bm25Ranker", () => {
 		expect(ranker.rank("tickets", documents, 8)).toEqual([]);
 	});
 
+	// https://github.com/earendil-works/pi/issues/11331
+	it("ranks a tool first when the query is its name", () => {
+		const named = [
+			tool("mcp__app__queue_read", "Read a queued message of this thread."),
+			tool(
+				"mcp__dev-app__thread_read",
+				"Read durable state and a paginated timeline from any thread. The default view returns user messages, assistant messages and plans. Continue with afterPosition until hasMore is false.",
+			),
+		].map((entry) => createToolSearchDocument(entry));
+		const ranker = new Bm25Ranker();
+		for (const query of [
+			"thread_read",
+			"dev_app__thread_read",
+			"mcp__dev-app__thread_read",
+			"Thread_Read",
+			"thread read",
+		]) {
+			expect(ranker.rank(query, named, 1).map((match) => match.name)).toEqual(["mcp__dev-app__thread_read"]);
+		}
+		// Partial names still rank by BM25.
+		expect(ranker.rank("read", named, 1).map((match) => match.name)).toEqual(["mcp__app__queue_read"]);
+	});
+
 	it("includes the namespace in the search text", () => {
 		const document = createToolSearchDocument(tool("mcp__x__run", "Run it."), {
 			name: "mcp__x",