@vanillagreen/pi-qol 1.4.3 → 1.6.0

package/extensions/qol.ts

@@ -2,7 +2,9 @@ import type { ExtensionAPI, ExtensionCommandContext, ExtensionContext } from "@e
 import type { TUI } from "@earendil-works/pi-tui";
 import { criticalInfo, lastAssistantTextFromAgentEnd, needsDirection, taskStats } from "./qol/agent-end.js";
 import { getQuestionService, readCavemanBridge, type QuestionOpenedEventLike } from "./qol/bridges.js";
+import { BudgetGuardDriver } from "./qol/budget-guard-runtime.js";
 import {
+	budgetGuardTrigger,
 	compactionTriggerReason,
 	handleQolBranchSummary,
 	handleQolCompaction,
@@ -44,6 +46,7 @@ import {
 	restorePendingQueueThemePatch,
 } from "./qol/pending-queue.js";
 import { permissionGateMatch, permissionGatePrompt } from "./qol/permission-gate.js";
+import { createRateLimitAutoResumeController, RATE_LIMIT_AUTO_RESUME_EVENT } from "./qol/rate-limit-auto-resume.js";
 import {
 	autoRenameEnabled,
 	autoRenameNotify,
@@ -53,6 +56,7 @@ import {
 	isStaleCtxError,
 	withAutoRenamePrefix,
 } from "./qol/session-rename.js";
+import { createScheduleController, getScheduleArgumentCompletions } from "./qol/schedule.js";
 import {
 	consumePendingSessionSearchContext,
 	openQolSessionSearch,
@@ -257,6 +261,35 @@ export default function qol(pi: ExtensionAPI): void {
 		idleCompactionTimer = undefined;
 	};
 
+	const budgetGuardDriver = new BudgetGuardDriver();
+
+	const resetBudgetGuard = () => {
+		budgetGuardDriver.reset();
+	};
+
+	const maybeFireBudgetGuard = (ctx: ExtensionContext) => {
+		let trigger;
+		try {
+			trigger = budgetGuardTrigger(ctx);
+		} catch (error) {
+			if (isStaleCtxError(error)) return;
+			throw error;
+		}
+		const notifySafely = (message: string, level: "info" | "warning" | "error") => {
+			try {
+				if (ctx.hasUI && settingBoolean("compaction.notify", true, ctx.cwd)) ctx.ui.notify(message, level);
+			} catch (error) {
+				if (isStaleCtxError(error)) return;
+				throw error;
+			}
+		};
+		budgetGuardDriver.dispatch({
+			compact: typeof ctx.compact === "function" ? ctx.compact.bind(ctx) : undefined,
+			notify: notifySafely,
+			trigger,
+		});
+	};
+
 	const scheduleIdleCompaction = (ctx: ExtensionContext) => {
 		clearIdleCompactionTimer();
 		if (!settingBoolean("compaction.idleEnabled", false, ctx.cwd)) return;
@@ -344,7 +377,13 @@ export default function qol(pi: ExtensionAPI): void {
 	let lastSessionTitle: string | undefined;
 
 	const requestRender = () => activeTui?.requestRender();
+	const scheduleController = createScheduleController(pi);
+	scheduleController.setOnChange(requestRender);
+	const rateLimitAutoResumeController = createRateLimitAutoResumeController(pi);
+	rateLimitAutoResumeController.setOnChange(requestRender);
+	const statuslineEnabled = (ctx: ExtensionContext): boolean => settingBoolean("statusline.enabled", true, ctx.cwd);
 	const refreshStatusline = (ctx: ExtensionContext) => {
+		if (!statuslineEnabled(ctx)) return Promise.resolve();
 		if (refreshInFlight) return refreshInFlight;
 		refreshInFlight = refreshGitState(pi, ctx)
 			.then((next) => {
@@ -503,13 +542,22 @@ export default function qol(pi: ExtensionAPI): void {
 		const event = data as QuestionOpenedEventLike;
 		notifyQuestionOpened(currentCtx, event, "question");
 	});
+	pi.events.on(RATE_LIMIT_AUTO_RESUME_EVENT, (payload: unknown) => {
+		if (!currentCtx) return;
+		rateLimitAutoResumeController.noteExternalRateLimitEvent(payload, currentCtx);
+	});
 
 	pi.on("session_start", (event, ctx) => {
 		currentCtx = ctx;
 		subscribeCavemanBridge();
 		latestSystemPromptOptions = undefined;
 		resetAutoRename();
+		resetBudgetGuard();
 		resetThinkingTimer(ctx);
+		const scheduleEnabled = settingBoolean("enableScheduleCommand", true, ctx.cwd);
+		const rateLimitAutoResumeEnabled = rateLimitAutoResumeController.enabled(ctx);
+		if (scheduleEnabled) scheduleController.restoreFromBranch(ctx);
+		else scheduleController.clearTimers();
 		void consumePendingSessionSearchContext(pi, ctx, event.reason);
 		installAutocompleteHintStyling(ctx);
 		installPendingQueueThemePatch(ctx);
@@ -517,7 +565,8 @@ export default function qol(pi: ExtensionAPI): void {
 			ctx.ui.setHiddenThinkingLabel(hiddenThinkingLabel(ctx.ui.theme, ctx.cwd));
 			applyWorkingIndicatorMode(ctx);
 			gitState = makeFallbackGitState(ctx.cwd);
-			void refreshStatusline(ctx);
+			const showStatusline = statuslineEnabled(ctx);
+			if (showStatusline) void refreshStatusline(ctx);
 			installSessionTitle(ctx);
 			installSessionTitleSync(ctx);
 			ctx.ui.setEditorComponent((tui, theme, keybindings) => {
@@ -526,27 +575,32 @@ export default function qol(pi: ExtensionAPI): void {
 					? new QolCompactPromptEditor(tui, theme, keybindings, Math.max(0, Math.floor(settingNumber("inputBottomPaddingLines", DEFAULT_INPUT_BOTTOM_PADDING_LINES, ctx.cwd))), ctx)
 					: new QolEditor(tui, theme, keybindings, ctx);
 			});
-			const statusWidgetTimer = setTimeout(() => {
-				ctx.ui.setWidget("statusline", (tui, theme) => {
-					activeTui = tui;
-					return {
-						invalidate() {},
-						render(width: number): string[] {
-							return [renderStatusLine(width, ctx, gitState ?? makeFallbackGitState(ctx.cwd), pi, theme)];
-						},
-					};
-				});
-			}, 0);
-			statusWidgetTimer.unref?.();
-			if (settingBoolean("replaceFooter", true, ctx.cwd)) {
-				ctx.ui.setFooter((tui, _theme, footerData) => {
-					activeTui = tui;
-					const unsubscribe = footerData.onBranchChange(() => {
-						void refreshStatusline(ctx);
-						requestRender();
+			if (showStatusline || scheduleEnabled || rateLimitAutoResumeEnabled) {
+				const statusWidgetTimer = setTimeout(() => {
+					ctx.ui.setWidget("statusline", (tui, theme) => {
+						activeTui = tui;
+						return {
+							invalidate() {},
+							render(width: number): string[] {
+								const lines = rateLimitAutoResumeController.enabled(ctx) ? rateLimitAutoResumeController.renderPreviewLines(width) : [];
+								if (scheduleEnabled) lines.push(...scheduleController.renderPreviewLines(width));
+								if (showStatusline) lines.push(renderStatusLine(width, ctx, gitState ?? makeFallbackGitState(ctx.cwd), pi, theme));
+								return lines;
+							},
+						};
 					});
-					return { dispose: unsubscribe, invalidate() {}, render: () => [] };
-				});
+				}, 0);
+				statusWidgetTimer.unref?.();
+				if (showStatusline && settingBoolean("replaceFooter", true, ctx.cwd)) {
+					ctx.ui.setFooter((tui, _theme, footerData) => {
+						activeTui = tui;
+						const unsubscribe = footerData.onBranchChange(() => {
+							void refreshStatusline(ctx);
+							requestRender();
+						});
+						return { dispose: unsubscribe, invalidate() {}, render: () => [] };
+					});
+				}
 			}
 		}
 		startQuestionSubscription(ctx);
@@ -568,7 +622,10 @@ export default function qol(pi: ExtensionAPI): void {
 	pi.on("session_shutdown", (_event, ctx) => {
 		cavemanUnsubscribe?.();
 		cavemanUnsubscribe = undefined;
+		scheduleController.clearTimers();
+		rateLimitAutoResumeController.clearTimers();
 		resetAutoRename();
+		resetBudgetGuard();
 		clearIdleCompactionTimer();
 		clearQuestionSubscribeTimer();
 		if (sessionSearchWarmupTimer) clearTimeout(sessionSearchWarmupTimer);
@@ -598,6 +655,7 @@ export default function qol(pi: ExtensionAPI): void {
 	pi.on("agent_start", (_event, ctx) => {
 		clearIdleCompactionTimer();
 		clearTmuxWindowMark();
+		rateLimitAutoResumeController.noteAgentStart(ctx);
 		if (ctx.hasUI) {
 			void refreshStatusline(ctx);
 			requestRender();
@@ -624,6 +682,7 @@ export default function qol(pi: ExtensionAPI): void {
 		}
 	});
 	pi.on("message_end", (event, ctx) => {
+		rateLimitAutoResumeController.noteMessageEnd(event, ctx);
 		if (!updateThinkingTimerEnabled(ctx)) return;
 		const message = event.message as any;
 		if (!message || message.role !== "assistant" || !Array.isArray(message.content)) return;
@@ -637,6 +696,8 @@ export default function qol(pi: ExtensionAPI): void {
 			void refreshStatusline(ctx);
 			requestRender();
 		}
+		rateLimitAutoResumeController.noteAgentEnd(event, ctx);
+		maybeFireBudgetGuard(ctx);
 		scheduleIdleCompaction(ctx);
 		void attemptAutoRename(ctx);
 		const text = lastAssistantTextFromAgentEnd(event, ctx);
@@ -660,10 +721,21 @@ export default function qol(pi: ExtensionAPI): void {
 		sendQolNotification(ctx, "ready", settingString("notification.readyMessage", "Ready for input", ctx.cwd), "info", "ready");
 	});
 	pi.on("session_compact", (_event, ctx) => {
+		// After a successful compaction usage drops below the budget, so reset the
+		// crossing key so the next threshold crossing re-fires the guard.
+		budgetGuardDriver.noteSessionCompacted();
 		if (!ctx.hasUI) return;
 		void refreshStatusline(ctx);
 		requestRender();
 	});
+	pi.on("session_tree", (_event, ctx) => {
+		currentCtx = ctx;
+		if (settingBoolean("enableScheduleCommand", true, ctx.cwd)) scheduleController.restoreFromBranch(ctx);
+		else scheduleController.clearTimers();
+	});
+	pi.on("after_provider_response", (event, ctx) => {
+		rateLimitAutoResumeController.noteProviderResponse(event, ctx);
+	});
 	pi.on("session_before_compact", (event, ctx) => handleQolCompaction(event, ctx));
 	pi.on("session_before_tree", (event, ctx) => handleQolBranchSummary(event, ctx));
 	pi.on("tool_call", async (event: any, ctx) => {
@@ -721,6 +793,14 @@ export default function qol(pi: ExtensionAPI): void {
 		});
 	}
 
+	if (settingBoolean("enableScheduleCommand", true)) {
+		pi.registerCommand("schedule", {
+			description: "Send a user message after a timer without invoking the model now.",
+			getArgumentCompletions: getScheduleArgumentCompletions,
+			handler: async (args, ctx) => scheduleController.handleCommand(args, ctx),
+		});
+	}
+
 	if (settingBoolean("enableContextCommand", true)) {
 		pi.registerMessageRenderer(CONTEXT_USAGE_MESSAGE_TYPE, renderQolContextUsageMessage);
 		pi.registerCommand("context", {

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@vanillagreen/pi-qol",
-  "version": "1.4.3",
-  "description": "Pi quality-of-life extension: compact statusline/π prompt, reliable multiline input, styled pasted-image chips, session naming/search/context import, handoff, permission prompts, notifications, custom compaction, and a collapsed-thinking timer.",
+  "version": "1.6.0",
+  "description": "Pi quality-of-life extension: compact statusline/π prompt, reliable multiline input, styled pasted-image chips, session naming/search/context import, scheduled prompts, handoff, permission prompts, notifications, custom compaction, and a collapsed-thinking timer.",
   "license": "MIT",
   "keywords": [
     "pi-package",
@@ -16,6 +16,9 @@
     ],
     "image": "https://raw.githubusercontent.com/vanillagreencom/vstack/main/pi-extensions/pi-qol/assets/settings-panel.png"
   },
+  "scripts": {
+    "test": "bun test ./tests"
+  },
   "vstack": {
     "extensionManager": {
       "displayName": "QOL",
@@ -30,10 +33,33 @@
           "apply": "reload",
           "requiresReload": true
         },
+        {
+          "key": "glyphStyle",
+          "label": "Glyph style",
+          "description": "Glyphs for vstack-owned UI chrome. Unicode is the default; ASCII uses plain box/status/separator characters for terminal compatibility. pi-tool-renderer's global override wins when set.",
+          "type": "enum",
+          "enumValues": [
+            "unicode",
+            "ascii"
+          ],
+          "default": "unicode",
+          "category": "Rendering",
+          "apply": "live"
+        },
+        {
+          "key": "statusline.enabled",
+          "label": "Show compact statusline",
+          "description": "Render the QOL statusline row with repo, branch, model, thinking level, and context percentage. Disable when another statusline/footer extension owns this UI.",
+          "type": "boolean",
+          "default": true,
+          "category": "Statusline",
+          "apply": "reload",
+          "requiresReload": true
+        },
         {
           "key": "replaceFooter",
           "label": "Replace built-in footer",
-          "description": "Hide Pi default footer while the compact QOL statusline is active.",
+          "description": "Hide Pi default footer while the compact QOL statusline is enabled.",
           "type": "boolean",
           "default": true,
           "category": "Statusline",
@@ -318,6 +344,61 @@
           "apply": "reload",
           "requiresReload": true
         },
+        {
+          "key": "enableScheduleCommand",
+          "label": "Enable /schedule command",
+          "description": "Register /schedule for timer-based user messages, e.g. /schedule 20m retry after rate reset. The timer runs in the extension without invoking the model immediately.",
+          "type": "boolean",
+          "default": true,
+          "category": "Session",
+          "apply": "reload",
+          "requiresReload": true
+        },
+        {
+          "key": "rateLimitAutoResume.enabled",
+          "label": "Auto-resume after rate limits",
+          "description": "When Pi detects a rate-limit stop with a reset time, schedule a follow-up user message at the reset time so the agent can continue if nothing else has continued since.",
+          "type": "boolean",
+          "default": false,
+          "category": "Rate Limit Recovery",
+          "apply": "live"
+        },
+        {
+          "key": "rateLimitAutoResume.message",
+          "label": "Auto-resume message",
+          "description": "User message sent after a detected rate limit resets.",
+          "type": "string",
+          "default": "Hit API rate limit. It has been reset. Continue",
+          "category": "Rate Limit Recovery",
+          "apply": "live"
+        },
+        {
+          "key": "rateLimitAutoResume.bufferSeconds",
+          "label": "Reset buffer seconds",
+          "description": "Seconds to wait after the reported reset time before sending the auto-resume message.",
+          "type": "number",
+          "default": 10,
+          "category": "Rate Limit Recovery",
+          "apply": "live"
+        },
+        {
+          "key": "rateLimitAutoResume.maxDelayMinutes",
+          "label": "Maximum auto-resume delay",
+          "description": "Ignore detected reset times farther away than this many minutes, to avoid stale or malformed schedules.",
+          "type": "number",
+          "default": 1440,
+          "category": "Rate Limit Recovery",
+          "apply": "live"
+        },
+        {
+          "key": "rateLimitAutoResume.notify",
+          "label": "Notify on auto-resume",
+          "description": "Show Pi UI notices when rate-limit auto-resume is scheduled, sent, or fails.",
+          "type": "boolean",
+          "default": true,
+          "category": "Rate Limit Recovery",
+          "apply": "live"
+        },
         {
           "key": "handoffReviewPrompt",
           "label": "Review handoff prompt",
@@ -654,9 +735,9 @@
         {
           "key": "compaction.model",
           "label": "Compaction model",
-          "description": "Summarizer model as provider/model, a bare model id, or current. Default mirrors Pi's custom-compaction example.",
+          "description": "Summarizer model as provider/model, a bare model id, or current. Default uses Pi's active model.",
           "type": "string",
-          "default": "google/gemini-2.5-flash",
+          "default": "current",
           "category": "Compaction",
           "apply": "live"
         },
@@ -782,6 +863,60 @@
           "category": "Compaction Threshold",
           "apply": "live"
         },
+        {
+          "key": "compaction.budgetGuardEnabled",
+          "label": "Long-session budget guard",
+          "description": "Trigger compaction on agent_end (no idle wait) when context usage crosses budgetPercent or budgetTokens. Designed for autonomous/Flightdeck runs so transcript growth cannot silently hit provider/buffer limits. Fires once per threshold crossing, not in a loop.",
+          "type": "boolean",
+          "default": true,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
+        {
+          "key": "compaction.budgetPercent",
+          "label": "Budget guard percent",
+          "description": "Context-window percentage that fires the budget guard at agent_end. Set -1 to disable percent-based firing and rely on absolute tokens instead.",
+          "type": "number",
+          "default": 85,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
+        {
+          "key": "compaction.budgetTokens",
+          "label": "Budget guard token limit",
+          "description": "Absolute token count above which the budget guard fires at agent_end. Set -1 to use only budgetPercent.",
+          "type": "number",
+          "default": -1,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
+        {
+          "key": "compaction.maxInputChars",
+          "label": "Chunked compaction input cap",
+          "description": "Maximum serialized characters sent to the summarizer in a single request. Inputs larger than this are split into chunks, summarized chunk-by-chunk, then merged via a summary-of-summaries pass so the compaction request payload itself cannot blow past provider buffer limits. Set 0 to disable chunking.",
+          "type": "number",
+          "default": 240000,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
+        {
+          "key": "compaction.handoffArtifactEnabled",
+          "label": "Write pre-compaction handoff artifact",
+          "description": "Before compaction, write a JSON handoff artifact under ~/.pi/agent/vstack/sessions/<session>/pi-qol/handoff/ containing the previous summary, last task state, and recently-referenced file/artifact paths. Always writes a stamped file and a latest.json pointer. Runs whenever the QOL bounded compaction handler runs — that is, on every budget-guard-triggered compaction (regardless of compaction.customEnabled), and on manual/idle compactions only when compaction.customEnabled is on. Write failures surface a QOL warning and a handoffArtifactError field in the compaction details.",
+          "type": "boolean",
+          "default": true,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
+        {
+          "key": "compaction.transcriptRiskWarnChars",
+          "label": "Transcript-risk warn budget (chars)",
+          "description": "Show a transcript-risk warning in /context when the serialized payload of the messages-to-send exceeds this many characters, even if the token count is still below the model context window. Set 0 to disable the warning.",
+          "type": "number",
+          "default": 600000,
+          "category": "Compaction Budget Guard",
+          "apply": "live"
+        },
         {
           "key": "thinkingLabel.text",
           "label": "Hidden thinking label",