@oh-my-pi/pi-coding-agent 4.2.0 → 4.2.1

package/CHANGELOG.md

@@ -2,6 +2,32 @@
 
 ## [Unreleased]
 
+## [4.2.1] - 2026-01-11
+### Added
+
+- Added automatic discovery and listing of AGENTS.md files in the system prompt, providing agents with an authoritative list of project-specific instruction files without runtime searching
+- Added `planner` built-in agent for comprehensive implementation planning with slow model
+
+### Changed
+
+- Refactored skill discovery to use unified `loadSkillsFromDir` helper across all providers, reducing code duplication
+- Updated skill discovery to scan only `skills/*/SKILL.md` entries instead of recursive walks in Codex provider
+- Added guidance to Task tool documentation to isolate file scopes when assigning tasks to prevent agent conflicts
+- Updated Task tool documentation to emphasize that subagents have no access to conversation history and require all relevant context to be explicitly passed
+- Revised task agent prompt to clarify that subagents have full tool access and can make file edits, run commands, and create files
+- OpenAI Codex: updated to use bundled system prompt from upstream
+- Changed `complete` tool to make `data` parameter optional when aborting, while still requiring it for successful completions
+- Skills discovery now scans only `skills/*/SKILL.md` entries instead of recursive walks
+
+### Removed
+
+- Removed `architect-plan`, `implement`, and `implement-with-critic` built-in agent commands
+
+### Fixed
+
+- Fixed editor border rendering glitch after canceling slash command autocomplete
+- Fixed login/logout credential path message to reference agent.db
+
 ## [4.2.0] - 2026-01-10
 
 ### Added

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-coding-agent",
-	"version": "4.2.0",
+	"version": "4.2.1",
 	"description": "Coding agent CLI with read, bash, edit, write tools and session management",
 	"type": "module",
 	"ompConfig": {
@@ -39,10 +39,10 @@
 		"prepublishOnly": "bun run generate-template && bun run clean && bun run build"
 	},
 	"dependencies": {
-		"@oh-my-pi/pi-ai": "4.2.0",
-		"@oh-my-pi/pi-agent-core": "4.2.0",
-		"@oh-my-pi/pi-git-tool": "4.2.0",
-		"@oh-my-pi/pi-tui": "4.2.0",
+		"@oh-my-pi/pi-ai": "4.2.1",
+		"@oh-my-pi/pi-agent-core": "4.2.1",
+		"@oh-my-pi/pi-git-tool": "4.2.1",
+		"@oh-my-pi/pi-tui": "4.2.1",
 		"@openai/agents": "^0.3.7",
 		"@sinclair/typebox": "^0.34.46",
 		"ajv": "^8.17.1",

package/src/core/agent-session.ts

@@ -17,7 +17,7 @@ import type { Agent, AgentEvent, AgentMessage, AgentState, AgentTool, ThinkingLe
 import type { AssistantMessage, ImageContent, Message, Model, TextContent, Usage } from "@oh-my-pi/pi-ai";
 import { isContextOverflow, modelsAreEqual, supportsXhigh } from "@oh-my-pi/pi-ai";
 import type { Rule } from "../capability/rule";
-import { getAuthPath } from "../config";
+import { getAgentDbPath } from "../config";
 import { theme } from "../modes/interactive/theme/theme";
 import { type BashResult, executeBash as executeBashCommand, executeBashWithOperations } from "./bash-executor";
 import {
@@ -761,7 +761,7 @@ export class AgentSession {
 		if (!this.model) {
 			throw new Error(
 				"No model selected.\n\n" +
-					`Use /login, set an API key environment variable, or create ${getAuthPath()}\n\n` +
+					`Use /login, set an API key environment variable, or create ${getAgentDbPath()}\n\n` +
 					"Then use /model to select a model.",
 			);
 		}
@@ -771,7 +771,7 @@ export class AgentSession {
 		if (!apiKey) {
 			throw new Error(
 				`No API key found for ${this.model.provider}.\n\n` +
-					`Use /login, set an API key environment variable, or create ${getAuthPath()}`,
+					`Use /login, set an API key environment variable, or create ${getAgentDbPath()}`,
 			);
 		}
 

package/src/core/export-html/index.ts

@@ -1,7 +1,6 @@
 import { existsSync, writeFileSync } from "node:fs";
 import { basename } from "node:path";
-import type { AgentState, AgentTool } from "@oh-my-pi/pi-agent-core";
-import { buildCodexPiBridge, getCodexInstructions } from "@oh-my-pi/pi-ai";
+import type { AgentState } from "@oh-my-pi/pi-agent-core";
 import { APP_NAME } from "../../config";
 import { getResolvedThemeColors, getThemeExportColors } from "../../modes/interactive/theme/theme";
 import { SessionManager } from "../session-manager";
@@ -14,33 +13,6 @@ export interface ExportOptions {
 	themeName?: string;
 }
 
-/** Info about Codex injection to show inline with model_change entries. */
-interface CodexInjectionInfo {
-	/** Codex instructions text. */
-	instructions: string;
-	/** Bridge text (tool list). */
-	bridge: string;
-}
-
-/** Build Codex injection info for display inline with model_change entries. */
-async function buildCodexInjectionInfo(tools?: AgentTool[]): Promise<CodexInjectionInfo | undefined> {
-	let instructions: string | null = null;
-	try {
-		instructions = await getCodexInstructions("gpt-5.1-codex");
-	} catch {
-		// Cache miss is expected before the first Codex request.
-	}
-
-	const bridgeText = buildCodexPiBridge(tools);
-	const instructionsText =
-		instructions ?? "(Codex instructions not cached. Run a Codex request to populate the local cache.)";
-
-	return {
-		instructions: instructionsText,
-		bridge: bridgeText,
-	};
-}
-
 /** Parse a color string to RGB values. */
 function parseColor(color: string): { r: number; g: number; b: number } | undefined {
 	const hexMatch = color.match(/^#([0-9a-fA-F]{2})([0-9a-fA-F]{2})([0-9a-fA-F]{2})$/);
@@ -125,8 +97,6 @@ interface SessionData {
 	entries: ReturnType<SessionManager["getEntries"]>;
 	leafId: string | null;
 	systemPrompt?: string;
-	/** Info for rendering Codex injection inline with model_change entries. */
-	codexInjectionInfo?: CodexInjectionInfo;
 	tools?: { name: string; description: string }[];
 }
 
@@ -158,7 +128,6 @@ export async function exportSessionToHtml(
 		entries: sm.getEntries(),
 		leafId: sm.getLeafId(),
 		systemPrompt: state?.systemPrompt,
-		codexInjectionInfo: await buildCodexInjectionInfo(state?.tools),
 		tools: state?.tools?.map((t) => ({ name: t.name, description: t.description })),
 	};
 
@@ -180,7 +149,6 @@ export async function exportFromFile(inputPath: string, options?: ExportOptions
 		header: sm.getHeader(),
 		entries: sm.getEntries(),
 		leafId: sm.getLeafId(),
-		codexInjectionInfo: await buildCodexInjectionInfo(),
 	};
 
 	const html = generateHtml(sessionData, opts.themeName);

package/src/core/system-prompt.ts

@@ -148,6 +148,45 @@ function stripQuotes(value: string): string {
 	return value.replace(/^"|"$/g, "");
 }
 
+const AGENTS_MD_PATTERN = "**/AGENTS.md";
+const AGENTS_MD_LIMIT = 200;
+
+interface AgentsMdSearch {
+	scopePath: string;
+	limit: number;
+	pattern: string;
+	files: string[];
+}
+
+function normalizePath(value: string): string {
+	return value.replace(/\\/g, "/");
+}
+
+function listAgentsMdFiles(root: string, limit: number): string[] {
+	try {
+		const entries = Array.from(
+			new Bun.Glob(AGENTS_MD_PATTERN).scanSync({ cwd: root, onlyFiles: true, dot: false, absolute: false }),
+		);
+		const normalized = entries
+			.map((entry) => normalizePath(entry))
+			.filter((entry) => entry.length > 0 && !entry.includes("node_modules"))
+			.sort();
+		return normalized.length > limit ? normalized.slice(0, limit) : normalized;
+	} catch {
+		return [];
+	}
+}
+
+function buildAgentsMdSearch(cwd: string): AgentsMdSearch {
+	const files = listAgentsMdFiles(cwd, AGENTS_MD_LIMIT);
+	return {
+		scopePath: ".",
+		limit: AGENTS_MD_LIMIT,
+		pattern: AGENTS_MD_PATTERN,
+		files,
+	};
+}
+
 function getOsName(): string {
 	switch (process.platform) {
 		case "win32":
@@ -625,6 +664,7 @@ export function buildSystemPrompt(options: BuildSystemPromptOptions = {}): strin
 
 	// Resolve context files: use provided or discover
 	const contextFiles = providedContextFiles ?? loadProjectContextFiles({ cwd: resolvedCwd });
+	const agentsMdSearch = buildAgentsMdSearch(resolvedCwd);
 
 	// Build tool descriptions array
 	// Priority: toolNames (explicit list) > tools (Map) > defaults
@@ -663,6 +703,7 @@ export function buildSystemPrompt(options: BuildSystemPromptOptions = {}): strin
 			customPrompt: resolvedCustomPrompt,
 			appendPrompt: resolvedAppendPrompt ?? "",
 			contextFiles,
+			agentsMdSearch,
 			toolDescriptions: toolDescriptionsArray,
 			git,
 			skills: filteredSkills,
@@ -678,6 +719,7 @@ export function buildSystemPrompt(options: BuildSystemPromptOptions = {}): strin
 		environment: getEnvironmentInfo(),
 		systemPromptCustomization: systemPromptCustomization ?? "",
 		contextFiles,
+		agentsMdSearch,
 		git,
 		skills: filteredSkills,
 		rules: rules ?? [],

package/src/core/tools/complete.ts

@@ -79,7 +79,7 @@ export function createCompleteTool(session: ToolSession) {
 		: Type.Any({ description: "Structured JSON output (no schema specified)" });
 
 	const completeParams = Type.Object({
-		data: dataSchema,
+		data: Type.Optional(dataSchema),
 		status: Type.Optional(
 			Type.Union([Type.Literal("success"), Type.Literal("aborted")], {
 				default: "success",
@@ -99,8 +99,11 @@ export function createCompleteTool(session: ToolSession) {
 		execute: async (_toolCallId, params) => {
 			const status = params.status ?? "success";
 
-			// Skip schema validation when aborting - the agent is giving up
+			// Skip validation when aborting - data is optional for aborts
 			if (status === "success") {
+				if (params.data === undefined) {
+					throw new Error("data is required when status is 'success'");
+				}
 				if (schemaError) {
 					throw new Error(`Invalid output schema: ${schemaError}`);
 				}

package/src/core/tools/task/commands.ts

@@ -9,17 +9,13 @@ import { type SlashCommand, slashCommandCapability } from "../../../capability/s
 import { loadSync } from "../../../discovery";
 
 // Embed command markdown files at build time
-import architectPlanMd from "../../../prompts/agents/architect-plan.md" with { type: "text" };
-import implementMd from "../../../prompts/agents/implement.md" with { type: "text" };
-import implementWithCriticMd from "../../../prompts/agents/implement-with-critic.md" with { type: "text" };
 import initMd from "../../../prompts/agents/init.md" with { type: "text" };
+import plannerMd from "../../../prompts/agents/planner.md" with { type: "text" };
 import { renderPromptTemplate } from "../../prompt-templates";
 
 const EMBEDDED_COMMANDS: { name: string; content: string }[] = [
-	{ name: "architect-plan.md", content: renderPromptTemplate(architectPlanMd) },
-	{ name: "implement-with-critic.md", content: renderPromptTemplate(implementWithCriticMd) },
-	{ name: "implement.md", content: renderPromptTemplate(implementMd) },
 	{ name: "init.md", content: renderPromptTemplate(initMd) },
+	{ name: "planner.md", content: renderPromptTemplate(plannerMd) },
 ];
 
 export const EMBEDDED_COMMAND_TEMPLATES: ReadonlyArray<{ name: string; content: string }> = EMBEDDED_COMMANDS;

package/src/discovery/builtin.ts

@@ -5,7 +5,7 @@
  * .pi is an alias for backwards compatibility.
  */
 
-import { basename, dirname, isAbsolute, join, resolve } from "path";
+import { dirname, isAbsolute, join, resolve } from "path";
 import { type ContextFile, contextFileCapability } from "../capability/context-file";
 import { type Extension, type ExtensionManifest, extensionCapability } from "../capability/extension";
 import { type ExtensionModule, extensionModuleCapability } from "../capability/extension-module";
@@ -16,7 +16,7 @@ import { type MCPServer, mcpCapability } from "../capability/mcp";
 import { type Prompt, promptCapability } from "../capability/prompt";
 import { type Rule, ruleCapability } from "../capability/rule";
 import { type Settings, settingsCapability } from "../capability/settings";
-import { type Skill, type SkillFrontmatter, skillCapability } from "../capability/skill";
+import { type Skill, skillCapability } from "../capability/skill";
 import { type SlashCommand, slashCommandCapability } from "../capability/slash-command";
 import { type SystemPrompt, systemPromptCapability } from "../capability/system-prompt";
 import { type CustomTool, toolCapability } from "../capability/tool";
@@ -27,6 +27,7 @@ import {
 	expandEnvVarsDeep,
 	getExtensionNameFromPath,
 	loadFilesFromDir,
+	loadSkillsFromDir,
 	parseFrontmatter,
 	parseJSON,
 	SOURCE_PATHS,
@@ -190,64 +191,18 @@ registerProvider<SystemPrompt>(systemPromptCapability.id, {
 });
 
 // Skills
-function loadSkillFromFile(ctx: LoadContext, path: string, level: "user" | "project"): Skill | null {
-	const content = ctx.fs.readFile(path);
-	if (!content) return null;
-
-	const { frontmatter, body } = parseFrontmatter(content);
-	const skillDir = dirname(path);
-	const parentDirName = basename(skillDir);
-	const name = (frontmatter.name as string) || parentDirName;
-
-	if (!frontmatter.description) return null;
-
-	return {
-		name,
-		path,
-		content: body,
-		frontmatter: frontmatter as SkillFrontmatter,
-		level,
-		_source: createSourceMeta(PROVIDER_ID, path, level),
-	};
-}
-
-function loadSkillsRecursive(ctx: LoadContext, dir: string, level: "user" | "project"): LoadResult<Skill> {
-	const items: Skill[] = [];
-	const warnings: string[] = [];
-
-	if (!ctx.fs.isDir(dir)) return { items, warnings };
-
-	for (const name of ctx.fs.readDir(dir)) {
-		if (name.startsWith(".") || name === "node_modules") continue;
-
-		const path = join(dir, name);
-
-		if (ctx.fs.isDir(path)) {
-			const skillFile = join(path, "SKILL.md");
-			if (ctx.fs.isFile(skillFile)) {
-				const skill = loadSkillFromFile(ctx, skillFile, level);
-				if (skill) items.push(skill);
-			}
-
-			const sub = loadSkillsRecursive(ctx, path, level);
-			items.push(...sub.items);
-			if (sub.warnings) warnings.push(...sub.warnings);
-		} else if (name === "SKILL.md") {
-			const skill = loadSkillFromFile(ctx, path, level);
-			if (skill) items.push(skill);
-		}
-	}
-
-	return { items, warnings };
-}
-
 function loadSkills(ctx: LoadContext): LoadResult<Skill> {
 	const items: Skill[] = [];
 	const warnings: string[] = [];
 
 	for (const { dir, level } of getConfigDirs(ctx)) {
 		const skillsDir = join(dir, "skills");
-		const result = loadSkillsRecursive(ctx, skillsDir, level);
+		const result = loadSkillsFromDir(ctx, {
+			dir: skillsDir,
+			providerId: PROVIDER_ID,
+			level,
+			requireDescription: true,
+		});
 		items.push(...result.items);
 		if (result.warnings) warnings.push(...result.warnings);
 	}

package/src/discovery/claude.ts

@@ -24,7 +24,7 @@ import {
 	expandEnvVarsDeep,
 	getExtensionNameFromPath,
 	loadFilesFromDir,
-	parseFrontmatter,
+	loadSkillsFromDir,
 	parseJSON,
 } from "./helpers";
 
@@ -218,78 +218,25 @@ function loadSkills(ctx: LoadContext): LoadResult<Skill> {
 	const items: Skill[] = [];
 	const warnings: string[] = [];
 
-	// User-level: ~/.claude/skills/*/SKILL.md
-	const userBase = getUserClaude(ctx);
-	const userSkillsDir = join(userBase, "skills");
-
-	if (ctx.fs.isDir(userSkillsDir)) {
-		const skillDirs = ctx.fs.readDir(userSkillsDir);
-
-		for (const dirName of skillDirs) {
-			if (dirName.startsWith(".")) continue;
-
-			const skillDir = join(userSkillsDir, dirName);
-			if (!ctx.fs.isDir(skillDir)) continue;
-
-			const skillFile = join(skillDir, "SKILL.md");
-			if (!ctx.fs.isFile(skillFile)) continue;
-
-			const content = ctx.fs.readFile(skillFile);
-			if (!content) {
-				warnings.push(`Failed to read ${skillFile}`);
-				continue;
-			}
-
-			const { frontmatter, body } = parseFrontmatter(content);
-			const name = (frontmatter.name as string) || dirName;
-
-			items.push({
-				name,
-				path: skillFile,
-				content: body,
-				frontmatter,
-				level: "user",
-				_source: createSourceMeta(PROVIDER_ID, skillFile, "user"),
-			});
-		}
-	}
+	const userSkillsDir = join(getUserClaude(ctx), "skills");
+	const userResult = loadSkillsFromDir(ctx, {
+		dir: userSkillsDir,
+		providerId: PROVIDER_ID,
+		level: "user",
+	});
+	items.push(...userResult.items);
+	if (userResult.warnings) warnings.push(...userResult.warnings);
 
-	// Project-level: <project>/.claude/skills/*/SKILL.md
 	const projectBase = getProjectClaude(ctx);
 	if (projectBase) {
 		const projectSkillsDir = join(projectBase, "skills");
-
-		if (ctx.fs.isDir(projectSkillsDir)) {
-			const skillDirs = ctx.fs.readDir(projectSkillsDir);
-
-			for (const dirName of skillDirs) {
-				if (dirName.startsWith(".")) continue;
-
-				const skillDir = join(projectSkillsDir, dirName);
-				if (!ctx.fs.isDir(skillDir)) continue;
-
-				const skillFile = join(skillDir, "SKILL.md");
-				if (!ctx.fs.isFile(skillFile)) continue;
-
-				const content = ctx.fs.readFile(skillFile);
-				if (!content) {
-					warnings.push(`Failed to read ${skillFile}`);
-					continue;
-				}
-
-				const { frontmatter, body } = parseFrontmatter(content);
-				const name = (frontmatter.name as string) || dirName;
-
-				items.push({
-					name,
-					path: skillFile,
-					content: body,
-					frontmatter,
-					level: "project",
-					_source: createSourceMeta(PROVIDER_ID, skillFile, "project"),
-				});
-			}
-		}
+		const projectResult = loadSkillsFromDir(ctx, {
+			dir: projectSkillsDir,
+			providerId: PROVIDER_ID,
+			level: "project",
+		});
+		items.push(...projectResult.items);
+		if (projectResult.warnings) warnings.push(...projectResult.warnings);
 	}
 
 	return { items, warnings };

package/src/discovery/codex.ts

@@ -33,6 +33,7 @@ import {
 	discoverExtensionModulePaths,
 	getExtensionNameFromPath,
 	loadFilesFromDir,
+	loadSkillsFromDir,
 	parseFrontmatter,
 	SOURCE_PATHS,
 } from "./helpers";
@@ -209,51 +210,25 @@ function loadSkills(ctx: LoadContext): LoadResult<Skill> {
 	const items: Skill[] = [];
 	const warnings: string[] = [];
 
-	// User level: ~/.codex/skills/
 	const userSkillsDir = join(ctx.home, SOURCE_PATHS.codex.userBase, "skills");
-	const userResult = loadFilesFromDir(ctx, userSkillsDir, PROVIDER_ID, "user", {
-		extensions: ["md"],
-		recursive: true,
-		transform: (name, content, path, source) => {
-			const { frontmatter, body } = parseFrontmatter(content);
-			const skillName = frontmatter.name || name.replace(/\.md$/, "");
-
-			return {
-				name: String(skillName),
-				path,
-				content: body,
-				frontmatter,
-				level: "user" as const,
-				_source: source,
-			};
-		},
+	const userResult = loadSkillsFromDir(ctx, {
+		dir: userSkillsDir,
+		providerId: PROVIDER_ID,
+		level: "user",
 	});
 	items.push(...userResult.items);
-	warnings.push(...(userResult.warnings || []));
+	if (userResult.warnings) warnings.push(...userResult.warnings);
 
-	// Project level: .codex/skills/
 	const codexDir = ctx.fs.walkUp(".codex", { dir: true });
 	if (codexDir) {
 		const projectSkillsDir = join(codexDir, "skills");
-		const projectResult = loadFilesFromDir(ctx, projectSkillsDir, PROVIDER_ID, "project", {
-			extensions: ["md"],
-			recursive: true,
-			transform: (name, content, path, source) => {
-				const { frontmatter, body } = parseFrontmatter(content);
-				const skillName = frontmatter.name || name.replace(/\.md$/, "");
-
-				return {
-					name: String(skillName),
-					path,
-					content: body,
-					frontmatter,
-					level: "project" as const,
-					_source: source,
-				};
-			},
+		const projectResult = loadSkillsFromDir(ctx, {
+			dir: projectSkillsDir,
+			providerId: PROVIDER_ID,
+			level: "project",
 		});
 		items.push(...projectResult.items);
-		warnings.push(...(projectResult.warnings || []));
+		if (projectResult.warnings) warnings.push(...projectResult.warnings);
 	}
 
 	return { items, warnings };

package/src/discovery/helpers.ts

@@ -4,6 +4,7 @@
 
 import { join, resolve } from "path";
 import { parse as parseYAML } from "yaml";
+import type { Skill, SkillFrontmatter } from "../capability/skill";
 import type { LoadContext, LoadResult, SourceMeta } from "../capability/types";
 
 /**
@@ -126,6 +127,56 @@ export function parseFrontmatter(content: string): {
 	}
 }
 
+export function loadSkillsFromDir(
+	ctx: LoadContext,
+	options: {
+		dir: string;
+		providerId: string;
+		level: "user" | "project";
+		requireDescription?: boolean;
+	},
+): LoadResult<Skill> {
+	const items: Skill[] = [];
+	const warnings: string[] = [];
+	const { dir, level, providerId, requireDescription = false } = options;
+
+	if (!ctx.fs.isDir(dir)) {
+		return { items, warnings };
+	}
+
+	for (const name of ctx.fs.readDir(dir)) {
+		if (name.startsWith(".") || name === "node_modules") continue;
+
+		const skillDir = join(dir, name);
+		if (!ctx.fs.isDir(skillDir)) continue;
+
+		const skillFile = join(skillDir, "SKILL.md");
+		if (!ctx.fs.isFile(skillFile)) continue;
+
+		const content = ctx.fs.readFile(skillFile);
+		if (!content) {
+			warnings.push(`Failed to read ${skillFile}`);
+			continue;
+		}
+
+		const { frontmatter, body } = parseFrontmatter(content);
+		if (requireDescription && !frontmatter.description) {
+			continue;
+		}
+
+		items.push({
+			name: (frontmatter.name as string) || name,
+			path: skillFile,
+			content: body,
+			frontmatter: frontmatter as SkillFrontmatter,
+			level,
+			_source: createSourceMeta(providerId, skillFile, level),
+		});
+	}
+
+	return { items, warnings };
+}
+
 /**
  * Expand environment variables in a string.
  * Supports ${VAR} and ${VAR:-default} syntax.
@@ -286,7 +337,7 @@ export function discoverExtensionModulePaths(ctx: LoadContext, dir: string): str
 	const discovered: string[] = [];
 
 	for (const name of ctx.fs.readDir(dir)) {
-		if (name.startsWith(".")) continue;
+		if (name.startsWith(".") || name === "node_modules") continue;
 
 		const entryPath = join(dir, name);
 

package/src/modes/interactive/controllers/selector-controller.ts

@@ -2,7 +2,7 @@ import type { ThinkingLevel } from "@oh-my-pi/pi-agent-core";
 import type { OAuthProvider } from "@oh-my-pi/pi-ai";
 import type { Component } from "@oh-my-pi/pi-tui";
 import { Input, Loader, Spacer, Text } from "@oh-my-pi/pi-tui";
-import { getAuthPath } from "../../../config";
+import { getAgentDbPath } from "../../../config";
 import { SessionManager } from "../../../core/session-manager";
 import { setPreferredImageProvider, setPreferredWebSearchProvider } from "../../../core/tools/index";
 import { disableProvider, enableProvider } from "../../../discovery";
@@ -546,7 +546,7 @@ export class SelectorController {
 								),
 							);
 							this.ctx.chatContainer.addChild(
-								new Text(theme.fg("dim", `Credentials saved to ${getAuthPath()}`), 1, 0),
+								new Text(theme.fg("dim", `Credentials saved to ${getAgentDbPath()}`), 1, 0),
 							);
 							this.ctx.ui.requestRender();
 						} catch (error: unknown) {
@@ -566,7 +566,7 @@ export class SelectorController {
 								),
 							);
 							this.ctx.chatContainer.addChild(
-								new Text(theme.fg("dim", `Credentials removed from ${getAuthPath()}`), 1, 0),
+								new Text(theme.fg("dim", `Credentials removed from ${getAgentDbPath()}`), 1, 0),
 							);
 							this.ctx.ui.requestRender();
 						} catch (error: unknown) {

package/src/modes/interactive/interactive-mode.ts

@@ -149,6 +149,9 @@ export class InteractiveMode implements InteractiveModeContext {
 		this.statusContainer = new Container();
 		this.editor = new CustomEditor(getEditorTheme());
 		this.editor.setUseTerminalCursor(true);
+		this.editor.onAutocompleteCancel = () => {
+			this.ui.requestRender(true);
+		};
 		try {
 			this.historyStorage = HistoryStorage.open();
 			this.editor.setHistoryStorage(this.historyStorage);

package/src/prompts/agents/planner.md

@@ -0,0 +1,112 @@
+---
+name: planner
+description: Software architect that explores codebase and produces detailed implementation plans
+tools: read, grep, find, ls, bash
+spawns: explore
+model: pi/slow, gpt-5.2-codex, gpt-5.2, codex, gpt
+---
+
+<role>Senior software architect producing implementation plans. READ-ONLY — no file modifications, no state changes.</role>
+
+<context>
+Another engineer will execute your plan without re-exploring the codebase. Your plan must be specific enough to implement directly.
+</context>
+
+<process>
+## Phase 1: Understand
+
+1. Parse the task requirements precisely
+2. Identify ambiguities — list assumptions you're making
+3. Spawn parallel `explore` agents if the task spans multiple areas
+
+## Phase 2: Explore
+
+Investigate thoroughly before designing:
+
+1. Find existing patterns via grep/find
+2. Read key files to understand current architecture
+3. Trace data flow through relevant code paths
+4. Identify types, interfaces, and contracts involved
+5. Note dependencies between components
+
+Spawn `explore` agents for independent search areas. Synthesize findings.
+
+## Phase 3: Design
+
+Create implementation approach:
+
+1. List concrete changes required (files, functions, types)
+2. Define the sequence — what depends on what
+3. Identify edge cases and error conditions
+4. Consider alternatives; justify your choice
+5. Note potential pitfalls or tricky parts
+
+## Phase 4: Produce Plan
+
+Write a plan another engineer can execute without re-exploring the codebase.
+</process>
+
+<example>
+## Summary
+What we're building and why (one paragraph).
+
+## Changes
+1. **`path/to/file.ts`** — What to change
+   - Specific modifications
+2. **`path/to/other.ts`** — ...
+
+## Sequence
+1. X (no dependencies)
+2. Y (depends on X)
+3. Z (integration)
+
+## Edge Cases
+- Case: How to handle
+
+## Verification
+- [ ] Test command or check
+- [ ] Expected behavior
+
+## Critical Files
+- `path/to/file.ts` (lines 50-120) — Why to read
+</example>
+
+<example>
+## Summary
+Add rate limiting to the API gateway to prevent abuse. Requires middleware insertion and Redis integration for distributed counter storage.
+
+## Changes
+1. **`src/middleware/rate-limit.ts`** — New file
+   - Create `RateLimitMiddleware` class using sliding window algorithm
+   - Accept `maxRequests`, `windowMs`, `keyGenerator` options
+2. **`src/gateway/index.ts`** — Wire middleware
+   - Import and register before auth middleware (line 45)
+3. **`src/config/redis.ts`** — Add rate limit key prefix
+   - Add `RATE_LIMIT_PREFIX` constant
+
+## Sequence
+1. `rate-limit.ts` (standalone, no deps)
+2. `redis.ts` (config only)
+3. `gateway/index.ts` (integration)
+
+## Edge Cases
+- Redis unavailable: fail open with warning log
+- IPv6 addresses: normalize before using as key
+
+## Verification
+- [ ] `curl -X GET localhost:3000/api/test` 100x rapidly → 429 after limit
+- [ ] Redis CLI: `KEYS rate:*` shows entries
+
+## Critical Files
+- `src/middleware/auth.ts` (lines 20-50) — Pattern to follow
+- `src/types/middleware.ts` — Interface to implement
+</example>
+
+<requirements>
+- Plan must be specific enough to implement without additional exploration
+- Include exact file paths and line ranges where relevant
+- Sequence must respect dependencies
+- Verification must be concrete and testable
+</requirements>
+
+Keep going until complete. This matters — get it right.

package/src/prompts/agents/task.md

@@ -1,14 +1,15 @@
-You are a worker agent for delegated tasks in an isolated context. Finish only the assigned work and return the minimum useful result.
+You are a worker agent for delegated tasks. You have FULL access to all tools (edit, write, bash, grep, read, etc.) - use them as needed to complete your task.
+
+Finish only the assigned work and return the minimum useful result.
 
 Principles:
 
+- You CAN and SHOULD make file edits, run commands, and create files when your task requires it.
 - Be concise. No filler, repetition, or tool transcripts.
-- If blocked, ask a single focused question; otherwise proceed autonomously.
 - Prefer narrow search (grep/find) then read only needed ranges.
 - Avoid full-file reads unless necessary.
-- NEVER create files unless absolutely required. Prefer edits to existing files.
+- Prefer edits to existing files over creating new ones.
 - NEVER create documentation files (\*.md) unless explicitly requested.
-- Any file paths in your response MUST be absolute.
 - When spawning subagents with the Task tool, include a 5-8 word user-facing description.
 - Include the smallest relevant code snippet when discussing code or config.
 - Follow the main agent's instructions.

package/src/prompts/system/system-prompt.md

@@ -137,6 +137,11 @@ Before reading any file:
 
 ## Project Integration
 - Follow AGENTS.md by scope: nearest file applies, deeper overrides higher.
+- Do not search for AGENTS.md during execution; use this list as authoritative.
+{{#if agentsMdSearch.files.length}}
+Relevant files are:
+{{#list agentsMdSearch.files join="\n"}}- {{this}}{{/list}}
+{{/if}}
 - Resolve blockers before yielding.
 </instructions>
 

package/src/prompts/tools/task.md

@@ -2,6 +2,13 @@ Launch a new agent to handle complex, multi-step tasks autonomously.
 
 The Task tool launches specialized agents (workers) that autonomously handle complex tasks. Each agent type has specific capabilities and tools available to it.
 
+**CRITICAL: Subagents have NO access to conversation history.** They only see:
+1. Their agent-specific system prompt
+2. The `context` string you provide
+3. The `task` string you provide
+
+If you discussed requirements, plans, schemas, or decisions with the user, you MUST include that information in `context`. Subagents cannot see prior messages - they start fresh with only what you explicitly pass them.
+
 ## Available Agents
 
 {{#list agents prefix="- " join="\n"}}
@@ -26,8 +33,9 @@ The Task tool launches specialized agents (workers) that autonomously handle com
 - **Minimize tool chatter**: Avoid repeating large context; use Output tool with output ids for full logs
 - **Structured completion**: If `output` is provided, subagents must call `complete` to finish
 - **Parallelize**: Launch multiple agents concurrently whenever possible
+- **Isolate file scopes**: Assign each task distinct files or directories so agents don't conflict
 - **Results are intermediate data**: Agent findings provide context for YOU to perform actual work. Do not treat agent reports as "task complete" signals.
-- **Stateless invocations**: Each agent runs autonomously and returns a single final message. Include all necessary context and specify exactly what information to return.
+- **Stateless invocations**: Subagents have zero memory of your conversation. Pass ALL relevant context: requirements discussed, decisions made, schemas agreed upon, file paths mentioned. If you reference something from earlier discussion without including it, the subagent will fail.
 - **Trust outputs**: Agent results should generally be trusted
 - **Clarify intent**: Tell the agent whether you expect code changes or just research (search, file reads, web fetches)
 - **Proactive use**: If an agent description says to use it proactively, do so without waiting for explicit user request
@@ -35,7 +43,7 @@ The Task tool launches specialized agents (workers) that autonomously handle com
 ## Parameters
 
 - `agent`: Agent type to use for all tasks
-- `context`: Shared context string prepended to all task prompts
+- `context`: **Required context from conversation** - include ALL relevant info: requirements, schemas, decisions, constraints. Subagents cannot see chat history.
 - `model`: (optional) Model override (fuzzy matching, e.g., "sonnet", "opus")
 - `tasks`: Array of `{id, task, description}` - tasks to run in parallel (max {{MAX_PARALLEL_TASKS}}, {{MAX_CONCURRENCY}} concurrent)
   - `id`: Short CamelCase identifier for display (max 20 chars, e.g., "SessionStore", "LspRefactor")
@@ -46,30 +54,28 @@ The Task tool launches specialized agents (workers) that autonomously handle com
 ## Example
 
 <example>
-user: "Extract all hardcoded strings for i18n"
-assistant: I'll scan UI components and return structured string locations for internationalization.
+user: "Looks good, execute the plan"
+assistant: I'll execute the refactoring plan.
 assistant: Uses the Task tool:
 {
-  "agent": "explore",
-  "context": "Find hardcoded user-facing strings (labels, messages, errors). Ignore logs, comments, and internal identifiers.",
+  "agent": "task",
+  "context": "Refactoring the auth module into separate concerns.\n\nPlan:\n1. AuthProvider - Extract React context and provider from src/auth/index.tsx\n2. AuthApi - Extract API calls to src/auth/api.ts, use existing fetchJson helper\n3. AuthTypes - Move types to src/auth/types.ts, re-export from index\n\nConstraints:\n- Preserve all existing exports from src/auth/index.tsx\n- Use project's fetchJson (src/utils/http.ts), don't use raw fetch\n- No new dependencies",
   "output": {
     "properties": {
-      "strings": {
-        "elements": {
-          "properties": {
-            "file": { "type": "string" },
-            "line": { "type": "uint32" },
-            "text": { "type": "string" },
-            "suggestedKey": { "type": "string" }
-          }
-        }
-      }
+      "summary": { "type": "string" },
+      "decisions": { "elements": { "type": "string" } },
+      "concerns": { "elements": { "type": "string" } }
     }
   },
   "tasks": [
-    { "id": "Forms", "task": "Scan src/components/forms/", "description": "Extract form strings" },
-    { "id": "Modals", "task": "Scan src/components/modals/", "description": "Extract modal strings" },
-    { "id": "Pages", "task": "Scan src/pages/", "description": "Extract page strings" }
+    { "id": "AuthProvider", "task": "Execute step 1: Extract AuthProvider and AuthContext", "description": "Extract React context" },
+    { "id": "AuthApi", "task": "Execute step 2: Extract API calls to api.ts", "description": "Extract API layer" },
+    { "id": "AuthTypes", "task": "Execute step 3: Move types to types.ts", "description": "Extract types" }
   ]
 }
 </example>
+
+Key points:
+- **Plan in context**: The full plan is written once; each task references its step without repeating shared constraints
+- **Parallel execution**: 3 agents run concurrently, each owning one step - no duplicated work
+- **Structured output**: JTD schema ensures consistent reporting across all agents

package/src/prompts/agents/architect-plan.md

@@ -1,10 +0,0 @@
----
-description: Explore gathers context, planner creates implementation plan (no implementation)
----
-
-Use the subagent tool with the chain parameter to execute this workflow:
-
-1. First, use the "explore" agent to find all code relevant to: $@
-2. Then, use the "planner" agent to create an implementation plan for "$@" using the context from the previous step (use {previous} placeholder)
-
-Execute this as a chain, passing output between steps via {previous}. Do NOT implement - just return the plan.

package/src/prompts/agents/implement-with-critic.md

@@ -1,11 +0,0 @@
----
-description: Task implements, reviewer reviews, task applies feedback
----
-
-Use the subagent tool with the chain parameter to execute this workflow:
-
-1. First, use the "task" agent to implement: $@
-2. Then, use the "reviewer" agent to review the implementation from the previous step (use {previous} placeholder)
-3. Finally, use the "task" agent to apply the feedback from the review (use {previous} placeholder)
-
-Execute this as a chain, passing output between steps via {previous}.

package/src/prompts/agents/implement.md

@@ -1,11 +0,0 @@
----
-description: Full implementation workflow - explore gathers context, planner creates plan, task implements
----
-
-Use the subagent tool with the chain parameter to execute this workflow:
-
-1. First, use the "explore" agent to find all code relevant to: $@
-2. Then, use the "planner" agent to create an implementation plan for "$@" using the context from the previous step (use {previous} placeholder)
-3. Finally, use the "task" agent to implement the plan from the previous step (use {previous} placeholder)
-
-Execute this as a chain, passing output between steps via {previous}.