context: place usage fix in unreleased

CHANGELOG.md

@@ -181,6 +181,13 @@ Docs: https://docs.openclaw.ai
 - LINE/markdown: preserve underscores inside Latin, Cyrillic, and CJK words when stripping markdown, while still removing standalone `_italic_` markers on the shared text-runtime path used by LINE and TTS. (#47465) Thanks @jackjin1997.
 - Agents/failover: make overloaded same-provider retry count and retry delay configurable via `auth.cooldowns`, default to one retry with no delay, and document the model-fallback behavior.
 - Ollama/model picker: include configured Ollama models in the opted-in non-PI-native model catalog path so Ollama onboarding shows available models directly after provider selection. (#55290) Thanks @Luckymingxuan.
+- Telegram/audio: transcode Telegram voice-note `.ogg` attachments before the local `whisper-cli` auto fallback runs, and keep mention-preflight transcription enabled in auto mode when `tools.media.audio` is unset.
+- Matrix/direct rooms: recover fresh auto-joined 1:1 DMs without eagerly persisting invite-only `m.direct` mappings, while keeping named, aliased, and explicitly configured rooms on the room path. (#58024) Thanks @gumadeiras.
+- TTS: Restore 3.28 schema compatibility and fallback observability. (#57953) Thanks @joshavant.
+- Telegram/forum topics: restore reply routing to the active topic and keep ACP `sessions_spawn(..., thread=true, mode="session")` bound to that same topic instead of falling back to root chat or losing follow-up routing. (#56060) Thanks @one27001.
+- Config/SecretRef + Control UI: harden SecretRef redaction round-trip restore, block unsafe raw fallback (force Form mode when raw is unavailable), and preflight submitted-config SecretRefs before config write RPC persistence. (#58044) Thanks @joshavant.
+- Config/Telegram: migrate removed `channels.telegram.groupMentionsOnly` into `channels.telegram.groups["*"].requireMention` on load so legacy configs no longer crash at startup. (#55336) thanks @jameslcowan.
+- `/context detail` now compares the tracked prompt estimate with cached context usage and surfaces untracked provider/runtime overhead when present. (#28391) thanks @ImLukeF.
 
 ## 2026.3.31-beta.1
 
@@ -2325,7 +2332,6 @@ Docs: https://docs.openclaw.ai
 - Diagnostics/Stuck session signal: add configurable stuck-session warning threshold via `diagnostics.stuckSessionWarnMs` (default 120000ms) to reduce false-positive warnings on long multi-tool turns. (#31032)
 - Agents/error classification: check billing errors before context overflow heuristics in the agent runner catch block so spend-limit and quota errors show the billing-specific message instead of being misclassified as "Context overflow: prompt too large". (#40409) Thanks @ademczuk.
 - Memory/MMR CJK tokenization: add Han, kana, and hangul tokens plus adjacent bigrams so memory-search reranking can detect overlap for CJK text instead of treating unrelated snippets as identical. (#29396) Thanks @buyitsydney.
-- `/context detail` now compares the tracked prompt estimate with cached context usage and surfaces untracked provider/runtime overhead when present. (#28391) thanks @ImLukeF.
 
 ## 2026.2.26
 

src/auto-reply/reply/commands-context-report.test.ts

@@ -8,6 +8,8 @@ function makeParams(
   options?: {
     omitBootstrapLimits?: boolean;
     contextTokens?: number | null;
+    totalTokens?: number | null;
+    totalTokensFresh?: boolean;
   },
 ): HandleCommandsParams {
   return {
@@ -25,7 +27,8 @@ function makeParams(
     resolvedThinkLevel: "off",
     resolvedReasoningLevel: "off",
     sessionEntry: {
-      totalTokens: 123,
+      totalTokens: options?.totalTokens ?? 123,
+      totalTokensFresh: options?.totalTokensFresh ?? true,
       inputTokens: 100,
       outputTokens: 23,
       systemPromptReport: {
@@ -96,18 +99,27 @@ describe("buildContextReply", () => {
   it("shows tracked estimate and cached context delta in detail output", async () => {
     const result = await buildContextReply(
       makeParams("/context detail", false, {
-        contextTokens: 900,
+        contextTokens: 8_192,
+        totalTokens: 900,
       }),
     );
     expect(result.text).toContain("Tracked prompt estimate: 1,020 chars (~255 tok)");
     expect(result.text).toContain("Actual context usage (cached): 900 tok");
     expect(result.text).toContain("Untracked provider/runtime overhead: ~645 tok");
+    expect(result.text).toContain("Session tokens (cached): 900 total / ctx=8,192");
   });
 
   it("shows estimate-only detail output when cached context usage is unavailable", async () => {
-    const result = await buildContextReply(makeParams("/context detail", false));
+    const result = await buildContextReply(
+      makeParams("/context detail", false, {
+        contextTokens: 8_192,
+        totalTokens: 900,
+        totalTokensFresh: false,
+      }),
+    );
     expect(result.text).toContain("Tracked prompt estimate: 1,020 chars (~255 tok)");
     expect(result.text).toContain("Actual context usage (cached): unavailable");
+    expect(result.text).toContain("Session tokens (cached): unknown / ctx=8,192");
     expect(result.text).not.toContain("~645 tok");
   });
 });

src/auto-reply/reply/commands-context-report.ts

@@ -4,7 +4,10 @@ import {
   resolveBootstrapTotalMaxChars,
 } from "../../agents/pi-embedded-helpers.js";
 import { buildSystemPromptReport } from "../../agents/system-prompt-report.js";
-import type { SessionSystemPromptReport } from "../../config/sessions/types.js";
+import {
+  resolveFreshSessionTotalTokens,
+  type SessionSystemPromptReport,
+} from "../../config/sessions/types.js";
 import { estimateTokensFromChars } from "../../utils/cjk-chars.js";
 import type { ReplyPayload } from "../types.js";
 import { resolveCommandsSystemPromptBundle } from "./commands-system-prompt.js";
@@ -93,8 +96,10 @@ export async function buildContextReply(params: HandleCommandsParams): Promise<R
   }
 
   const report = await resolveContextReport(params);
+  const cachedContextUsageTokens = resolveFreshSessionTotalTokens(params.sessionEntry);
   const session = {
     totalTokens: params.sessionEntry?.totalTokens ?? null,
+    totalTokensFresh: params.sessionEntry?.totalTokensFresh ?? null,
     inputTokens: params.sessionEntry?.inputTokens ?? null,
     outputTokens: params.sessionEntry?.outputTokens ?? null,
     contextTokens: params.contextTokens ?? null,
@@ -188,8 +193,8 @@ export async function buildContextReply(params: HandleCommandsParams): Promise<R
       : [];
 
   const totalsLine =
-    session.totalTokens != null
-      ? `Session tokens (cached): ${formatInt(session.totalTokens)} total / ctx=${session.contextTokens ?? "?"}`
+    cachedContextUsageTokens != null
+      ? `Session tokens (cached): ${formatInt(cachedContextUsageTokens)} total / ctx=${session.contextTokens ?? "?"}`
       : `Session tokens (cached): unknown / ctx=${session.contextTokens ?? "?"}`;
   const sharedContextLines = [
     `Workspace: ${workspaceLabel}`,
@@ -228,14 +233,15 @@ export async function buildContextReply(params: HandleCommandsParams): Promise<R
     // `systemPrompt.chars` already includes injected files, skills, and tool-list text.
     // Add only tool schemas here so the tracked estimate stays disjoint.
     const trackedPromptChars = report.systemPrompt.chars + report.tools.schemaChars;
-    const trackedPromptTokens = estimateTokensFromChars(trackedPromptChars);
     const trackedPromptLine = `Tracked prompt estimate: ${formatCharsAndTokens(trackedPromptChars)}`;
     const actualContextLine =
-      session.contextTokens != null
-        ? `Actual context usage (cached): ${formatInt(session.contextTokens)} tok`
+      cachedContextUsageTokens != null
+        ? `Actual context usage (cached): ${formatInt(cachedContextUsageTokens)} tok`
         : "Actual context usage (cached): unavailable";
     const overheadTokens =
-      session.contextTokens != null ? session.contextTokens - trackedPromptTokens : null;
+      cachedContextUsageTokens != null
+        ? cachedContextUsageTokens - estimateTokensFromChars(trackedPromptChars)
+        : null;
     const overheadLine =
       overheadTokens == null
         ? null