pi-cache-optimizer 2.5.1 → 2.5.3

package/index.ts

@@ -4,6 +4,39 @@ import { homedir } from "node:os";
 import { dirname, join } from "node:path";
 import type { BuildSystemPromptOptions, ExtensionAPI, ExtensionContext } from "@earendil-works/pi-coding-agent";
 
+type MutableEnv = Record<string, string | undefined>;
+
+type CacheRetentionEnvSnapshot = {
+  wasSet: boolean;
+  value?: string;
+};
+
+const PI_CACHE_RETENTION_ENV = "PI_CACHE_RETENTION";
+const LONG_CACHE_RETENTION_VALUE = "long";
+
+function captureCacheRetentionEnv(env: MutableEnv = process.env): CacheRetentionEnvSnapshot {
+  return {
+    wasSet: Object.prototype.hasOwnProperty.call(env, PI_CACHE_RETENTION_ENV),
+    value: env[PI_CACHE_RETENTION_ENV],
+  };
+}
+
+function requestLongCacheRetention(env: MutableEnv = process.env): void {
+  if (!env[PI_CACHE_RETENTION_ENV] || env[PI_CACHE_RETENTION_ENV] !== LONG_CACHE_RETENTION_VALUE) {
+    env[PI_CACHE_RETENTION_ENV] = LONG_CACHE_RETENTION_VALUE;
+  }
+}
+
+function restoreCacheRetentionEnv(snapshot: CacheRetentionEnvSnapshot, env: MutableEnv = process.env): void {
+  if (snapshot.wasSet) {
+    env[PI_CACHE_RETENTION_ENV] = snapshot.value;
+  } else {
+    delete env[PI_CACHE_RETENTION_ENV];
+  }
+}
+
+const STARTUP_CACHE_RETENTION_ENV = captureCacheRetentionEnv();
+
 /**
  * Pi Cache Optimizer (formerly pi-deepseek-cache-optimizer)
  *
@@ -19,10 +52,9 @@ import type { BuildSystemPromptOptions, ExtensionAPI, ExtensionContext } from "@
 
 // ============================================================
 // Automatically request long prompt-cache retention when Pi supports it.
+// /cache-optimizer disable restores the startup value for this Pi process.
 // ============================================================
-if (!process.env.PI_CACHE_RETENTION || process.env.PI_CACHE_RETENTION !== "long") {
-  process.env.PI_CACHE_RETENTION = "long";
-}
+requestLongCacheRetention();
 
 type PiModel = NonNullable<ExtensionContext["model"]>;
 type UnknownRecord = Record<string, unknown>;
@@ -40,6 +72,8 @@ const OPENAI_PROMPT_CACHE_KEY_MAX_LENGTH = 64;
 const NO_SKILL_COMPRESSION_ENV = "PI_CACHE_OPTIMIZER_NO_SKILL_COMPRESSION";
 const NO_PROMPT_REWRITE_ENV = "PI_CACHE_OPTIMIZER_NO_PROMPT_REWRITE";
 
+let runtimeOptimizerEnabled = true;
+
 // WORM-flag: if optimizeSystemPrompt ever detects that its blind-replace
 // logic has accidentally truncated a structural marker (any XML tag or
 // HTML comment boundary marker present in the original prompt), we flip
@@ -657,11 +691,46 @@ function isDisabledEnv(value: string | undefined): boolean {
 }
 
 function shouldInjectOpenAIPromptCacheKey(): boolean {
+  if (!runtimeOptimizerEnabled) return false;
   if (isEnabledEnv(process.env[NO_OPENAI_CACHE_KEY_ENV])) return false;
   if (isDisabledEnv(process.env[OPENAI_CACHE_KEY_ENV])) return false;
   return true;
 }
 
+function setRuntimeOptimizerEnabled(enabled: boolean, env: MutableEnv = process.env): void {
+  runtimeOptimizerEnabled = enabled;
+  if (enabled) {
+    requestLongCacheRetention(env);
+  } else {
+    restoreCacheRetentionEnv(STARTUP_CACHE_RETENTION_ENV, env);
+  }
+}
+
+function isRuntimeOptimizerEnabled(): boolean {
+  return runtimeOptimizerEnabled;
+}
+
+function getOptimizerRuntimeModeLines(): string[] {
+  const state = runtimeOptimizerEnabled ? "enabled" : "disabled";
+  const lines: string[] = [];
+  lines.push(`Runtime state: ${state}`);
+  lines.push(`• Prompt rewrite: ${runtimeOptimizerEnabled && !isEnabledEnv(process.env[NO_PROMPT_REWRITE_ENV]) ? "on" : "off"}`);
+  lines.push(`• OpenAI prompt_cache_key fallback: ${shouldInjectOpenAIPromptCacheKey() ? "on" : "off"}`);
+  lines.push(`• Footer cache stats: on${runtimeOptimizerEnabled ? "" : " (comparison mode)"}`);
+  lines.push(`• Compat warnings: ${runtimeOptimizerEnabled ? "on" : "off"}`);
+  lines.push(`• ${PI_CACHE_RETENTION_ENV}: ${process.env[PI_CACHE_RETENTION_ENV] ?? "(unset)"}`);
+  if (!runtimeOptimizerEnabled) {
+    lines.push("This is a current-process switch. Run /reload or restart Pi to return to startup behavior.");
+  } else if (isEnabledEnv(process.env[NO_PROMPT_REWRITE_ENV]) || !shouldInjectOpenAIPromptCacheKey()) {
+    lines.push("Some features are still disabled by environment variables.");
+  }
+  return lines;
+}
+
+function formatOptimizerRuntimeMode(): string {
+  return getOptimizerRuntimeModeLines().join("\n");
+}
+
 function isAssistantMessage(message: unknown): boolean {
   return asRecord(message)?.role === "assistant";
 }
@@ -1443,6 +1512,40 @@ function describeMissingOpenAICompatibleProxyCompat(model: PiModel): string[] {
   return missing;
 }
 
+function buildSafeOpenAIProxyCompatSuggestion(missing: string[]): Record<string, boolean> {
+  const suggestion: Record<string, boolean> = {};
+  if (missing.includes("sendSessionAffinityHeaders")) {
+    suggestion.sendSessionAffinityHeaders = true;
+  }
+  return suggestion;
+}
+
+function getPromptCacheRetentionUnsupportedHint(): string {
+  return "If this channel returns `400 Unsupported parameter: prompt_cache_retention`, remove/avoid `supportsLongCacheRetention`; this extension does not write that field directly, but Pi may send it when long retention is requested and compat says the proxy supports it.";
+}
+
+function appendOpenAIProxyCompatAdviceLines(lines: string[], missing: string[], options: { includeJsonIntro?: boolean } = {}): void {
+  const suggestion = buildSafeOpenAIProxyCompatSuggestion(missing);
+  const hasSafeSuggestion = Object.keys(suggestion).length > 0;
+
+  if (hasSafeSuggestion) {
+    if (options.includeJsonIntro !== false) {
+      lines.push("Safe default suggestion:");
+    }
+    lines.push(JSON.stringify(suggestion, null, 2));
+  } else if (missing.includes("supportsLongCacheRetention")) {
+    lines.push("No safe automatic JSON change is recommended for `supportsLongCacheRetention`.");
+  }
+
+  if (missing.includes("sendSessionAffinityHeaders")) {
+    lines.push("- sendSessionAffinityHeaders: recommended for third-party proxies when supported; it helps keep one Pi session on the same upstream/backend.");
+  }
+  if (missing.includes("supportsLongCacheRetention")) {
+    lines.push("- supportsLongCacheRetention: optional. Enable only after your endpoint/proxy explicitly supports OpenAI long prompt cache retention.");
+    lines.push(`- ${getPromptCacheRetentionUnsupportedHint()}`);
+  }
+}
+
 /**
  * Build the warning text displayed to users when an OpenAI-family third-party
  * proxy is missing one or more cache/session-affinity compat flags.
@@ -1456,11 +1559,6 @@ function describeMissingOpenAICompatibleProxyCompat(model: PiModel): string[] {
  * exercise it via __internals_for_tests.
  */
 function buildOpenAIProxyCompatWarningText(key: string, missing: string[]): string {
-  const suggestion: Record<string, boolean> = {};
-  for (const flag of missing) {
-    suggestion[flag] = true;
-  }
-
   // Extract provider id from the model key (e.g. "otokapi/gpt-5.5" -> "otokapi").
   // If no slash is found, fall back to the key itself.
   const slashIdx = key.indexOf("/");
@@ -1469,19 +1567,11 @@ function buildOpenAIProxyCompatWarningText(key: string, missing: string[]): stri
   const modelsJsonPath = getModelsJsonDisplayPath();
   const lines: string[] = [
     `💡 pi-cache-optimizer: ${key} is a third-party GPT/OpenAI-compatible proxy but merged compat lacks ${missing.join(" and ")}.`,
-    `Edit ${modelsJsonPath} -> providers["${providerLabel}"] -> compat (at the same level as baseUrl/api/apiKey/models):`,
-    ``,
-    JSON.stringify(suggestion, null, 2),
+    `Edit ${modelsJsonPath} -> providers["${providerLabel}"] -> compat (at the same level as baseUrl/api/apiKey/models).`,
     ``,
   ];
 
-  for (const flag of missing) {
-    if (flag === "supportsLongCacheRetention") {
-      lines.push("- supportsLongCacheRetention: confirm your endpoint or proxy supports long prompt cache retention.");
-    } else if (flag === "sendSessionAffinityHeaders") {
-      lines.push("- sendSessionAffinityHeaders: keeps requests on the same backend for proxy cache locality (session affinity).");
-    }
-  }
+  appendOpenAIProxyCompatAdviceLines(lines, missing);
 
   return lines.join("\n");
 }
@@ -3070,7 +3160,10 @@ function describeRouterChannelDiagnostics(model: PiModel): string[] {
       "   • Return cache usage fields (prompt_cache_hit_tokens, etc.) in the response.",
     );
     notes.push(
-      `   Example compat: { "sendSessionAffinityHeaders": true, "supportsLongCacheRetention": true }`,
+      `   Safe compat default: { "sendSessionAffinityHeaders": true }`,
+    );
+    notes.push(
+      `   Add supportsLongCacheRetention only if the proxy explicitly supports prompt_cache_retention.`,
     );
 
     return notes;
@@ -3103,7 +3196,7 @@ function describeRouterChannelDiagnostics(model: PiModel): string[] {
   return notes;
 }
 
-function buildDoctorDiagnosis(model: PiModel): string {
+function buildDoctorDiagnosis(model: PiModel, options: { promptCacheRetention400?: boolean } = {}): string {
   const lines: string[] = [];
   lines.push(`Provider: ${model.provider}`);
   lines.push(`Model:    ${model.id}`);
@@ -3120,16 +3213,25 @@ function buildDoctorDiagnosis(model: PiModel): string {
     const key = modelKey(model);
     const slashIdx = key.indexOf("/");
     const providerLabel = slashIdx > 0 ? key.slice(0, slashIdx) : key;
-    const suggestion = Object.fromEntries(missing.map((f) => [f, true]));
     const modelsJsonPath = getModelsJsonDisplayPath();
-    lines.push(`Edit ${modelsJsonPath} -> providers["${providerLabel}"] -> compat (same level as baseUrl/api/apiKey/models):`);
-    lines.push(JSON.stringify(suggestion, null, 2));
+    lines.push(`Edit ${modelsJsonPath} -> providers["${providerLabel}"] -> compat (same level as baseUrl/api/apiKey/models).`);
+    appendOpenAIProxyCompatAdviceLines(lines, missing);
   } else if (isCompatCheckApplicable(model)) {
     lines.push("✅ Compat fully configured.");
   } else {
     lines.push("ℹ️ Compat check not applicable for this model.");
   }
 
+  if (isCompatCheckApplicable(model) && compat.supportsLongCacheRetention === true) {
+    lines.push("");
+    if (options.promptCacheRetention400) {
+      lines.push("⚠️  A 400 response was observed while supportsLongCacheRetention is enabled.");
+      lines.push(`   ${getPromptCacheRetentionUnsupportedHint()}`);
+    } else {
+      lines.push(`ℹ️ Long retention is enabled. ${getPromptCacheRetentionUnsupportedHint()}`);
+    }
+  }
+
   // ── Router/channel diagnostics ──
   const routerNotes = describeRouterChannelDiagnostics(model);
   if (routerNotes.length > 0) {
@@ -3280,22 +3382,22 @@ function buildCompatDiagnosis(model: PiModel): string | undefined {
   if (missing.length > 0) {
     const slashIdx = key.indexOf("/");
     const providerLabel = slashIdx > 0 ? key.slice(0, slashIdx) : key;
-    const suggestion = Object.fromEntries(missing.map((f) => [f, true]));
     const modelsJsonPath = getModelsJsonDisplayPath();
     lines.push(`Active model: ${key}`);
     lines.push(`Missing: ${missing.join(", ")}`);
     lines.push("");
     lines.push(`Edit ${modelsJsonPath} -> providers["${providerLabel}"] -> compat`);
-    lines.push(`(at the same level as baseUrl/api/apiKey/models) and add:`);
-    lines.push(JSON.stringify(suggestion, null, 2));
-    lines.push("");
-    lines.push(`Only enable if your endpoint supports them.`);
+    lines.push(`(at the same level as baseUrl/api/apiKey/models).`);
+    appendOpenAIProxyCompatAdviceLines(lines, missing);
   }
 
   // When compat is fully configured but router notes exist, prefix the status.
   if (routerNotes.length > 0 && missing.length === 0) {
     if (isCompatCheckApplicable(model)) {
       lines.push("✅ Compat fully configured.");
+      if (getCompat(model).supportsLongCacheRetention === true) {
+        lines.push(getPromptCacheRetentionUnsupportedHint());
+      }
     } else {
       lines.push("ℹ️ Compat check not applicable for this model.");
     }
@@ -3339,6 +3441,8 @@ export const __internals_for_tests = {
   isOpenAIFamilyToken,
   describeMissingOpenAIFamilyProxyCompat,
   describeMissingOpenAICompatibleProxyCompat,
+  buildSafeOpenAIProxyCompatSuggestion,
+  getPromptCacheRetentionUnsupportedHint,
   isOfficialOpenAIBaseUrl,
   // Non-GPT OpenAI-compatible model detection
   isKimiLikeModel,
@@ -3454,6 +3558,15 @@ export const __internals_for_tests = {
   modelKey,
   // Platform-friendly path helper
   getModelsJsonDisplayPath,
+  captureCacheRetentionEnv,
+  requestLongCacheRetention,
+  restoreCacheRetentionEnv,
+  setRuntimeOptimizerEnabled,
+  isRuntimeOptimizerEnabled,
+  getOptimizerRuntimeModeLines,
+  formatOptimizerRuntimeMode,
+  PI_CACHE_RETENTION_ENV,
+  LONG_CACHE_RETENTION_VALUE,
   // Integrity diagnostics
   getLastPromptIntegrityWarningAt,
   // Diagnostic command helpers
@@ -3492,6 +3605,8 @@ export const __internals_for_tests = {
 
 export default function (pi: ExtensionAPI) {
   const warnedModels = new Set<string>();
+  const promptCacheRetention400Models = new Set<string>();
+  const warnedPromptCacheRetention400Models = new Set<string>();
   let cacheStatsByModel: Record<string, CacheStats> = {};
   let cacheStatsLegacyFamily: Partial<Record<CacheProviderId, CacheStats>> = emptyAllCacheStats();
   let lastStatusText: string | undefined;
@@ -3582,6 +3697,17 @@ export default function (pi: ExtensionAPI) {
     return created;
   }
 
+  function resetCurrentSessionStats(): void {
+    const prefix = `${currentSessionHash || "_nosession"}:`;
+    for (const key of Object.keys(cacheStatsByModel)) {
+      if (key.startsWith(prefix)) delete cacheStatsByModel[key];
+    }
+    for (const key of Array.from(recentSamplesByModelKey.keys())) {
+      if (key.startsWith(prefix)) recentSamplesByModelKey.delete(key);
+    }
+    lastStatusText = undefined;
+  }
+
   async function persistCacheStats(ctx?: ExtensionContext): Promise<void> {
     try {
       await writePersistedCacheStats(getCacheStatsState(), currentSessionHashSet ? currentSessionHash : undefined);
@@ -3745,7 +3871,8 @@ export default function (pi: ExtensionAPI) {
       // cacheStatsByModel[sessionModelKey(model)] on first use.
       const sk = model ? sessionModelKey(model) : undefined;
       const stats = sk ? cacheStatsByModel[sk] : undefined;
-      statusText = formatCacheStats(adapter, stats ?? emptyCacheStats());
+      const statsText = formatCacheStats(adapter, stats ?? emptyCacheStats());
+      statusText = runtimeOptimizerEnabled ? statsText : `Cache Optimizer disabled · ${statsText}`;
     }
 
     // If optimizeSystemPrompt detected structural truncation on this or
@@ -3779,7 +3906,7 @@ export default function (pi: ExtensionAPI) {
     // update so the marker persists through stats changes and day
     // rollovers. Redundant setStatus calls are blocked by the
     // `lastStatusText` early return above.
-    if (statusText !== undefined && model) {
+    if (runtimeOptimizerEnabled && statusText !== undefined && model) {
       const compatMissing = describeMissingOpenAICompatibleProxyCompat(model);
       if (compatMissing.length > 0) {
         statusText = statusText + " ⚠️ compat";
@@ -3794,12 +3921,12 @@ export default function (pi: ExtensionAPI) {
 
   pi.on("session_start", async (event, ctx) => {
     await restoreCacheStats(event.reason, ctx);
-    notifyCacheCompatIfNeeded(ctx.model, ctx, warnedModels);
+    if (runtimeOptimizerEnabled) notifyCacheCompatIfNeeded(ctx.model, ctx, warnedModels);
     await publishStatus(ctx);
   });
 
   pi.on("model_select", async (event, ctx) => {
-    notifyCacheCompatIfNeeded(event.model, ctx, warnedModels);
+    if (runtimeOptimizerEnabled) notifyCacheCompatIfNeeded(event.model, ctx, warnedModels);
     await publishStatus(ctx, event.model);
   });
 
@@ -3838,6 +3965,8 @@ export default function (pi: ExtensionAPI) {
       }
     }
 
+    if (!runtimeOptimizerEnabled) return {};
+
     // Global opt-out: PI_CACHE_OPTIMIZER_NO_PROMPT_REWRITE=1 bypasses all
     // prompt mutations below (session-overview churn strip, skill compression,
     // and stable-prefix reordering). Footer stats and the OpenAI
@@ -3894,6 +4023,25 @@ export default function (pi: ExtensionAPI) {
     return addOpenAIPromptCacheKey(event.payload, getSessionPromptCacheKey(ctx));
   });
 
+  pi.on("after_provider_response", (event, ctx) => {
+    const model = ctx.model;
+    if (!runtimeOptimizerEnabled || !model) return;
+    if (event.status !== 400) return;
+    if (!isCompatCheckApplicable(model)) return;
+    if (getCompat(model).supportsLongCacheRetention !== true) return;
+
+    const key = modelKey(model);
+    promptCacheRetention400Models.add(key);
+    if (warnedPromptCacheRetention400Models.has(key)) return;
+    warnedPromptCacheRetention400Models.add(key);
+    ctx.ui.notify(
+      `⚠️ ${LOG_PREFIX}: ${key} returned HTTP 400 while supportsLongCacheRetention is enabled. ` +
+      getPromptCacheRetentionUnsupportedHint() +
+      ` Run /cache-optimizer doctor for the exact edit location.`,
+      "warning",
+    );
+  });
+
   pi.on("message_end", async (event, ctx) => {
     const adapter = selectAdapterForAssistantMessage(event.message, ctx.model);
     if (!adapter) return;
@@ -3929,6 +4077,8 @@ export default function (pi: ExtensionAPI) {
   // ────────────────────────────────────────────────────────────────
   // Register /cache-optimizer command
   // Subcommands:
+  //   enable  — enable runtime prompt/cache optimizations for this process
+  //   disable — disable runtime prompt/cache optimizations for this process
   //   doctor  — show current model/provider/api/baseUrl/compat status
   //             with low-hit diagnosis
   //   stats   — show active model stats bucket, recent trend, usage
@@ -3942,12 +4092,24 @@ export default function (pi: ExtensionAPI) {
       const model = cmdCtx.model;
       const subcommand = args.trim().toLowerCase().split(/\s+/)[0] || "help";
 
-      if (subcommand === "doctor") {
+      if (subcommand === "enable") {
+        setRuntimeOptimizerEnabled(true);
+        resetCurrentSessionStats();
+        await flushPersistCacheStats(cmdCtx as unknown as ExtensionContext);
+        await publishStatus(cmdCtx as unknown as ExtensionContext, model);
+        cmdCtx.ui.notify(`✅ Pi Cache Optimizer enabled for this Pi process. Current-session stats were reset for before/after comparison.\n${formatOptimizerRuntimeMode()}`, "info");
+      } else if (subcommand === "disable") {
+        setRuntimeOptimizerEnabled(false);
+        resetCurrentSessionStats();
+        await flushPersistCacheStats(cmdCtx as unknown as ExtensionContext);
+        await publishStatus(cmdCtx as unknown as ExtensionContext, model);
+        cmdCtx.ui.notify(`⏸️ Pi Cache Optimizer disabled for this Pi process. Current-session stats were reset and will keep collecting while disabled for comparison.\n${formatOptimizerRuntimeMode()}`, "warning");
+      } else if (subcommand === "doctor") {
         if (!model) {
           cmdCtx.ui.notify("No active model selected. Select a model first with /model or pi --model.", "warning");
           return;
         }
-        const diagnosis = buildDoctorDiagnosis(model);
+        const diagnosis = buildDoctorDiagnosis(model, { promptCacheRetention400: promptCacheRetention400Models.has(modelKey(model)) });
         const adapter = selectAdapterForModel(model);
         const sk = model ? sessionModelKey(model) : undefined;
         const statsState = sk ? cacheStatsByModel[sk] : undefined;
@@ -4020,18 +4182,32 @@ export default function (pi: ExtensionAPI) {
         // Try interactive selection menu when UI supports it
         if (cmdCtx.hasUI) {
           const menuOptions = [
-            "🩺 Doctor — Show current model cache configuration",
-            "📊 Stats — Show active model stats bucket and trend",
-            "⚙️  Compat — Show compat suggestion with edit instructions",
-            "🔄 Reset — Reset local session stats for current model",
-            "❌ Cancel",
+            "Enable — Turn on runtime optimizations",
+            "Disable — Turn off runtime optimizations",
+            "Doctor — Show cache configuration",
+            "Stats — Show cache stats and trend",
+            "Compat — Show compat suggestion",
+            "Reset — Reset local session stats",
+            "Cancel",
           ];
           const choice = await cmdCtx.ui.select("Cache Optimizer", menuOptions);
           if (choice === menuOptions[0]) {
+            setRuntimeOptimizerEnabled(true);
+            resetCurrentSessionStats();
+            await flushPersistCacheStats(cmdCtx as unknown as ExtensionContext);
+            await publishStatus(cmdCtx as unknown as ExtensionContext, model);
+            cmdCtx.ui.notify(`✅ Pi Cache Optimizer enabled for this Pi process. Current-session stats were reset for before/after comparison.\n${formatOptimizerRuntimeMode()}`, "info");
+          } else if (choice === menuOptions[1]) {
+            setRuntimeOptimizerEnabled(false);
+            resetCurrentSessionStats();
+            await flushPersistCacheStats(cmdCtx as unknown as ExtensionContext);
+            await publishStatus(cmdCtx as unknown as ExtensionContext, model);
+            cmdCtx.ui.notify(`⏸️ Pi Cache Optimizer disabled for this Pi process. Current-session stats were reset and will keep collecting while disabled for comparison.\n${formatOptimizerRuntimeMode()}`, "warning");
+          } else if (choice === menuOptions[2]) {
             if (!model) {
               cmdCtx.ui.notify("No active model selected. Select a model first with /model or pi --model.", "warning");
             } else {
-              const diagnosis = buildDoctorDiagnosis(model);
+              const diagnosis = buildDoctorDiagnosis(model, { promptCacheRetention400: promptCacheRetention400Models.has(modelKey(model)) });
               const adapter = selectAdapterForModel(model);
               const sk = model ? sessionModelKey(model) : undefined;
               const statsState = sk ? cacheStatsByModel[sk] : undefined;
@@ -4042,7 +4218,7 @@ export default function (pi: ExtensionAPI) {
                 : diagnosis;
               cmdCtx.ui.notify(fullDiagnosis, "info");
             }
-          } else if (choice === menuOptions[1]) {
+          } else if (choice === menuOptions[3]) {
             if (!model) {
               cmdCtx.ui.notify("No active model selected. Select a model first with /model or pi --model.", "warning");
             } else {
@@ -4053,7 +4229,7 @@ export default function (pi: ExtensionAPI) {
               const output = buildStatsOutput(model, adapter, statsState, samples);
               cmdCtx.ui.notify(output, "info");
             }
-          } else if (choice === menuOptions[2]) {
+          } else if (choice === menuOptions[4]) {
             if (!model) {
               cmdCtx.ui.notify("No active model selected. Select a model first with /model or pi --model.", "warning");
             } else {
@@ -4069,7 +4245,7 @@ export default function (pi: ExtensionAPI) {
                 );
               }
             }
-          } else if (choice === menuOptions[3]) {
+          } else if (choice === menuOptions[5]) {
             if (!model) {
               cmdCtx.ui.notify("No active model selected. Select a model first with /model or pi --model.", "warning");
             } else {
@@ -4098,11 +4274,15 @@ export default function (pi: ExtensionAPI) {
         // Fallback: text help when no interactive UI
         const diagnosis: string[] = [];
         diagnosis.push("📋 /cache-optimizer commands:");
+        diagnosis.push("  enable  — Enable prompt/cache optimizations for this Pi process");
+        diagnosis.push("  disable — Disable prompt/cache optimizations for this Pi process");
         diagnosis.push("  doctor  — Show current model/provider/api/baseUrl/compat and low-hit diagnosis");
         diagnosis.push("  stats   — Show active model stats bucket and recent trend");
         diagnosis.push("  compat  — Show compat suggestion with edit location");
         diagnosis.push("  reset   — Reset local session stats for current model (does not affect upstream)");
         diagnosis.push("");
+        diagnosis.push(formatOptimizerRuntimeMode());
+        diagnosis.push("");
         if (model) {
           const displayKey = modelKey(model);
           const missing = describeMissingOpenAICompatibleProxyCompat(model);

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "pi-cache-optimizer",
-  "version": "2.5.1",
-  "description": "Pi extension that improves provider-side KV/prompt cache hit rates (DeepSeek, OpenAI, Claude, Gemini) by reordering the system prompt, requesting long retention, and showing footer cache stats. Renamed from pi-deepseek-cache-optimizer.",
+  "version": "2.5.3",
+  "description": "Improve Pi prompt/KV cache hit rates with stable prompts, OpenAI-compatible cache keys, proxy compat warnings, and footer cache stats.",
   "keywords": [
     "pi-package",
     "deepseek",