openclaw · vincentkoc · Jun 24, 2026 · Jun 24, 2026 · Jun 24, 2026 · Jun 24, 2026
diff --git a/.github/workflows/sandbox-common-smoke.yml b/.github/workflows/sandbox-common-smoke.yml
@@ -57,11 +57,10 @@ jobs:
           BASE_IMAGE="openclaw-sandbox-smoke-base:bookworm-slim" \
             TARGET_IMAGE="openclaw-sandbox-common-smoke:bookworm-slim" \
             PACKAGES="ca-certificates" \
-            INSTALL_PNPM=0 \
             INSTALL_BUN=0 \
             INSTALL_BREW=0 \
             FINAL_USER=sandbox \
             scripts/sandbox-common-setup.sh
 
-          u="$(timeout --kill-after=30s 2m docker run --rm openclaw-sandbox-common-smoke:bookworm-slim sh -lc 'id -un')"
-          test "$u" = "sandbox"
+          timeout --kill-after=30s 2m docker run --rm openclaw-sandbox-common-smoke:bookworm-slim sh -lc \
+            'set -e; test "$(id -un)" = sandbox; node --version; pnpm --version'
diff --git a/docs/.generated/config-baseline.sha256 b/docs/.generated/config-baseline.sha256
@@ -1,4 +1,4 @@
-9246475f5771612a5fd12de38b153783c4a4cbb8b2682a5c40115916661c90f2  config-baseline.json
-6349131baaa1828f2a071f42e4d7b17c8966c59b6588c8a4c1a32ea5ea4dcd5e  config-baseline.core.json
+f5a5855ddd7aa8c23a732f257eceaa20fd163b1d5f342c909f4aef15aa8643cf  config-baseline.json
+b8dffdb1a328aaf728a0707ab04d21c65f1a225a2360042e10832aa608699716  config-baseline.core.json
 671979e86e4c4f59415d0a20879e838f9bbd883b3d29eeb02cb5131db8d187fe  config-baseline.channel.json
 94529978588d6e3776a86780b22cf9ff46a6f9957f2f178d3829403fad451ca7  config-baseline.plugin.json
diff --git a/docs/concepts/usage-tracking.md b/docs/concepts/usage-tracking.md
@@ -30,6 +30,68 @@ title: "Usage tracking"
 - CLI: `openclaw channels list` prints the same usage snapshot alongside provider config (use `--no-usage` to skip).
 - macOS menu bar: "Usage" section under Context (only if available).
 
+## Default usage footer mode
+
+`/usage off|tokens|full` sets the footer for a session and is remembered for that
+session. `messages.responseUsage` seeds that mode for sessions that have not
+chosen one, so the footer can be on by default without typing `/usage` each time.
+
+Set one mode for every channel, or a per-channel map with a `default` fallback:
+
+```jsonc
+{
+  "messages": {
+    "responseUsage": "tokens",
+    // or: { "default": "off", "discord": "full" }
+  },
+}
+```
+
+### Three distinct session states
+
+A session's `responseUsage` field has three representable states, each with
+different semantics:
+
+| State               | Stored value                    | Effective mode                                                        |
+| ------------------- | ------------------------------- | --------------------------------------------------------------------- |
+| **Unset / inherit** | `undefined` (absent)            | Falls through to `messages.responseUsage` config default, then `off`. |
+| **Explicit off**    | `"off"` (stored)                | Always off — a non-off config default cannot re-enable the footer.    |
+| **Explicit on**     | `"tokens"` or `"full"` (stored) | That mode, regardless of config default.                              |
+
+### Precedence
+
+Effective mode = session override → channel config entry → `default` → `off`.
+
+An explicit `/usage off` is **persisted** as the literal value `"off"` in the
+session, not the same as "unset." This means a non-off `messages.responseUsage`
+default cannot turn the footer back on once the user has explicitly disabled it.
+
+### Resetting vs. turning off
+
+- `/usage off` — forces the footer off and persists that choice. A configured
+  non-off default cannot override this.
+- `/usage reset` (aliases: `inherit`, `clear`, `default`) — clears the session
+  override. The session then **inherits** the effective config default
+  (`messages.responseUsage`). If no default is configured, the footer is off
+  (unchanged from before). Use this to "go back to default" without explicitly
+  turning the footer on.
+- A full session reset (`/reset` or `/new`) or a session rollover **preserves**
+  the explicit usage-mode preference so the user's display choice survives
+  session rollovers. Only `/usage reset` (and its aliases) actually clears the
+  override.
+
+### Toggle behavior
+
+`/usage` with no arguments cycles: off → tokens → full → off. The starting point
+for the cycle is the **effective** current mode (session override falling through
+to the config default when unset), so the cycle is always consistent with what
+the user sees in the footer.
+
+### Config
+
+With no config the prior behavior holds (footer off until `/usage`). Use
+`/usage reset` to clear a session override and re-inherit the configured default.
+
 ## Custom `/usage full` footer
 
 `/usage full` shows a built-in compact footer with model, reasoning, fast/slow,

@@ -415,7 +415,7 @@ If you installed OpenClaw via `npm install -g openclaw`, use the inline `docker
 
   </Step>
   <Step title="Optional: build the common image">
-    For a more functional sandbox image with common tooling (for example `curl`, `jq`, `nodejs`, `python3`, `git`):
+    For a more functional sandbox image with common tooling (for example `curl`, `jq`, Node 24, pnpm, `python3`, and `git`):
 
     From a source checkout:
 

diff --git a/docs/reference/token-use.md b/docs/reference/token-use.md
@@ -76,6 +76,8 @@ Use these in chat:
   configured for the active model.
 - `/usage off|tokens|full` → appends a **per-response usage footer** to every reply.
   - Persists per session (stored as `responseUsage`).
+  - `/usage reset` (aliases: `inherit`, `clear`, `default`) — clears the session
+    override so the session re-inherits the configured default.
   - `/usage full` shows estimated cost only when OpenClaw has usage metadata and
     local pricing for the active model. Otherwise it shows tokens only.
 - `/usage cost` → shows a local cost summary from OpenClaw session logs.

diff --git a/docs/tools/slash-commands.md b/docs/tools/slash-commands.md
@@ -240,7 +240,7 @@ plugins.
     | `/tasks` | List active/recent background tasks for the current session |
     | `/context [list\|detail\|map\|json]` | Explain how context is assembled |
     | `/whoami` | Show your sender id. Alias: `/id` |
-    | `/usage off\|tokens\|full\|cost` | Control the per-response usage footer or print a local cost summary |
+    | `/usage off\|tokens\|full\|reset\|cost` | Control the per-response usage footer (`reset`/`inherit`/`clear`/`default` clears the session override to re-inherit the configured default) or print a local cost summary |
   </Accordion>
 
   <Accordion title="Skills, allowlists, approvals">

diff --git a/docs/web/tui.md b/docs/web/tui.md
@@ -126,7 +126,7 @@ Session controls:
 - `/verbose <on|full|off>`
 - `/trace <on|off>`
 - `/reasoning <on|off|stream>`
-- `/usage <off|tokens|full>`
+- `/usage <off|tokens|full|reset>` (`reset`/`inherit`/`clear`/`default` clears the session override)
 - `/goal [status] | /goal start <objective> | /goal pause|resume|complete|block|clear`
 - `/elevated <on|off|ask|full>` (alias: `/elev`)
 - `/activation <mention|always>`

diff --git a/extensions/duckduckgo/src/ddg-client.ts b/extensions/duckduckgo/src/ddg-client.ts
@@ -36,21 +36,49 @@ type DuckDuckGoResult = {
 };
 
 function decodeHtmlEntities(text: string): string {
-  return text
-    .replace(/&amp;/g, "&")
-    .replace(/&lt;/g, "<")
-    .replace(/&gt;/g, ">")
-    .replace(/&quot;/g, '"')
-    .replace(/&apos;/g, "'")
-    .replace(/&#39;/g, "'")
-    .replace(/&#x27;/g, "'")
-    .replace(/&#x2F;/g, "/")
-    .replace(/&nbsp;/g, " ")
-    .replace(/&ndash;/g, "-")
-    .replace(/&mdash;/g, "--")
-    .replace(/&hellip;/g, "...")
-    .replace(/&#(\d+);/g, (_, code) => String.fromCodePoint(Number(code)))
-    .replace(/&#x([0-9a-f]+);/gi, (_, code) => String.fromCodePoint(Number.parseInt(code, 16)));
+  return text.replace(
+    /&(?:lt|gt|quot|apos|#39|#x27|#x2F|nbsp|ndash|mdash|hellip|amp|#\d+|#x[0-9a-f]+);/gi,
+    (entity) => {
+      const normalized = entity.toLowerCase();
+      if (normalized === "&lt;") {
+        return "<";
+      }
+      if (normalized === "&gt;") {
+        return ">";
+      }
+      if (normalized === "&quot;") {
+        return '"';
+      }
+      if (normalized === "&apos;" || normalized === "&#39;" || normalized === "&#x27;") {
+        return "'";
+      }
+      if (normalized === "&#x2f;") {
+        return "/";
+      }
+      if (normalized === "&nbsp;") {
+        return " ";
+      }
+      if (normalized === "&ndash;") {
+        return "-";
+      }
+      if (normalized === "&mdash;") {
+        return "--";
+      }
+      if (normalized === "&hellip;") {
+        return "...";
+      }
+      if (normalized === "&amp;") {
+        return "&";
+      }
+      if (normalized.startsWith("&#x")) {
+        return String.fromCodePoint(Number.parseInt(normalized.slice(3, -1), 16));
+      }
+      if (normalized.startsWith("&#")) {
+        return String.fromCodePoint(Number.parseInt(normalized.slice(2, -1), 10));
+      }
+      return entity;
+    },
+  );
 }
 
 function stripHtml(html: string): string {

diff --git a/extensions/duckduckgo/src/ddg-search-provider.test.ts b/extensions/duckduckgo/src/ddg-search-provider.test.ts
@@ -186,6 +186,17 @@ describe("duckduckgo web search provider", () => {
     );
   });
 
+  it("does not double-decode escaped entities (decodes &amp; last)", () => {
+    // A result whose text literally shows "&lt;" arrives double-encoded as
+    // "&amp;lt;". Decoding &amp; first would re-decode it into "<", corrupting
+    // the snippet; &amp; must be decoded last.
+    expect(ddgClientTesting.decodeHtmlEntities("How to escape &amp;lt; in HTML")).toBe(
+      "How to escape &lt; in HTML",
+    );
+    expect(ddgClientTesting.decodeHtmlEntities("a&amp;#39;b")).toBe("a&#39;b");
+    expect(ddgClientTesting.decodeHtmlEntities("a&#x26;amp;b")).toBe("a&amp;b");
+  });
+
   it("parses results when href appears before class", () => {
     const html = `
       <a href="https://duckduckgo.com/l/?uddg=https%3A%2F%2Fexample.com" class="result__a">

diff --git a/extensions/exa/src/exa-web-search-provider.runtime.ts b/extensions/exa/src/exa-web-search-provider.runtime.ts
@@ -20,6 +20,7 @@ import {
   wrapWebContent,
   writeCachedSearchPayload,
 } from "openclaw/plugin-sdk/provider-web-search";
+import { readResponseWithLimit } from "openclaw/plugin-sdk/response-limit-runtime";
 import {
   normalizeOptionalLowercaseString,
   normalizeOptionalString,
@@ -30,6 +31,10 @@ const EXA_SEARCH_TYPES = ["auto", "neural", "fast", "deep", "deep-reasoning", "i
 const EXA_FRESHNESS_VALUES = ["day", "week", "month", "year"] as const;
 const EXA_MAX_SEARCH_COUNT = 100;
 const EXA_ERROR_BODY_LIMIT_BYTES = 8 * 1024;
+// Exa search responses are untrusted external bodies. Cap the success JSON the
+// same way other bundled providers do (16 MiB) so a misbehaving or hostile
+// endpoint cannot stream an unbounded body into memory before we parse it.
+const EXA_SEARCH_JSON_MAX_BYTES = 16 * 1024 * 1024;
 
 type ExaConfig = {
   apiKey?: string;
@@ -70,9 +75,17 @@ type ExaSearchResponse = {
   results?: unknown;
 };
 
-async function readExaSearchResults(response: Response): Promise<ExaSearchResult[]> {
+async function readExaSearchResults(
+  response: Response,
+  opts?: { maxBytes?: number },
+): Promise<ExaSearchResult[]> {
+  const maxBytes = opts?.maxBytes ?? EXA_SEARCH_JSON_MAX_BYTES;
+  const bytes = await readResponseWithLimit(response, maxBytes, {
+    onOverflow: ({ maxBytes: maxBytesLocal }) =>
+      new Error(`Exa API response exceeds ${maxBytesLocal} bytes`),
+  });
   try {
-    return normalizeExaResults(await response.json());
+    return normalizeExaResults(JSON.parse(new TextDecoder().decode(bytes)));
   } catch (cause) {
     throw new Error("Exa API returned malformed JSON", { cause });
   }

diff --git a/extensions/exa/src/exa-web-search-provider.test.ts b/extensions/exa/src/exa-web-search-provider.test.ts
@@ -26,6 +26,33 @@ function cancelTrackedResponse(
   };
 }
 
+function streamingJsonResponse(params: { chunkCount: number; chunkSize: number }): {
+  response: Response;
+  getReadCount: () => number;
+} {
+  // Streaming fixture proves an oversized success body stops being read before
+  // the whole payload is buffered into memory.
+  let reads = 0;
+  const encoder = new TextEncoder();
+  const stream = new ReadableStream<Uint8Array>({
+    pull(controller) {
+      if (reads >= params.chunkCount) {
+        controller.close();
+        return;
+      }
+      reads += 1;
+      controller.enqueue(encoder.encode("a".repeat(params.chunkSize)));
+    },
+  });
+  return {
+    response: new Response(stream, {
+      status: 200,
+      headers: { "content-type": "application/json" },
+    }),
+    getReadCount: () => reads,
+  };
+}
+
 describe("exa web search provider", () => {
   it("exposes the expected metadata and selection wiring", () => {
     const provider = createExaWebSearchProvider();
@@ -265,6 +292,27 @@ describe("exa web search provider", () => {
     );
   });
 
+  it("parses well-formed Exa search JSON under the byte cap", async () => {
+    const response = new Response(
+      JSON.stringify({ results: [{ url: "https://example.com", title: "Example" }] }),
+      { status: 200, headers: { "content-type": "application/json" } },
+    );
+
+    await expect(testing.readExaSearchResults(response)).resolves.toEqual([
+      { url: "https://example.com", title: "Example" },
+    ]);
+  });
+
+  it("caps oversized Exa search JSON instead of buffering the whole body", async () => {
+    const streamed = streamingJsonResponse({ chunkCount: 64, chunkSize: 1024 });
+
+    await expect(
+      testing.readExaSearchResults(streamed.response, { maxBytes: 4096 }),
+    ).rejects.toThrow(/Exa API response exceeds 4096 bytes/);
+
+    expect(streamed.getReadCount()).toBeLessThan(64);
+  });
+
   it("bounds Exa API error bodies without using response.text()", async () => {
     const tracked = cancelTrackedResponse(`${"exa upstream unavailable ".repeat(1024)}tail`, {
       status: 503,

diff --git a/extensions/lmstudio/src/models.test.ts b/extensions/lmstudio/src/models.test.ts
@@ -59,21 +59,6 @@ describe("lmstudio-models", () => {
     }
     return JSON.parse(init.body) as unknown;
   };
-  // The model fetch/load helpers now read bodies through the shared byte-capped
-  // reader, so success-path mocks must be real Response objects with a body
-  // stream rather than bare `{ ok, json }` placeholders.
-  const jsonResponse = (payload: unknown, init?: ResponseInit): Response =>
-    new Response(JSON.stringify(payload), {
-      status: 200,
-      headers: { "content-type": "application/json" },
-      ...init,
-    });
-  const malformedJsonResponse = (init?: ResponseInit): Response =>
-    new Response("{ this is not valid json", {
-      status: 200,
-      headers: { "content-type": "application/json" },
-      ...init,
-    });
   const cancelTrackedResponse = (
     text: string,
     init: ResponseInit,
@@ -104,7 +89,6 @@ describe("lmstudio-models", () => {
   }) =>
     vi.fn(async (url: string | URL, _init?: RequestInit) => {
       const key = params?.key ?? "qwen3-8b-instruct";
-      void init;
       if (String(url).endsWith("/api/v1/models")) {
         return jsonResponse({
           models: [

diff --git a/extensions/memory-core/src/concept-vocabulary.test.ts b/extensions/memory-core/src/concept-vocabulary.test.ts
@@ -29,6 +29,30 @@ describe("concept vocabulary", () => {
     expect(tags).not.toContain("2026-04-04.md");
   });
 
+  it("preserves short protected-glossary terms past the latin minimum-length gate", () => {
+    const tags = deriveConceptTags({
+      path: "memory/2026-04-04.md",
+      snippet: "Store the session in kv and back up to s3 nightly.",
+    });
+
+    // "kv" and "s3" are 2-char latin glossary entries that the generic min-length-3 gate would drop.
+    expect(tags).toContain("kv");
+    expect(tags).toContain("s3");
+  });
+
+  it("does not surface short glossary terms that only appear inside longer words", () => {
+    const tags = deriveConceptTags({
+      path: "memory/2026-04-04.md",
+      snippet: "Played the mkv recording and tuned the css3 layout.",
+    });
+
+    // "kv"/"s3" are substrings of "mkv"/"css3"; whole-word matching must not emit them as tags.
+    expect(tags).not.toContain("kv");
+    expect(tags).not.toContain("s3");
+    expect(tags).toContain("mkv");
+    expect(tags).toContain("css3");
+  });
+
   it("extracts protected and segmented CJK concept tags", () => {
     const tags = deriveConceptTags({
       path: "memory/2026-04-04.md",