@shuyhere/bb-agent 0.0.14 → 0.0.16

package/CHANGELOG.md

@@ -5,12 +5,58 @@ All notable changes to BB-Agent will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+## [0.0.16] - 2026-04-15
+
+### Added
+
+- added `bb setup browser`, including platform-specific guidance and optional shell-profile persistence for configuring `BB_BROWSER`
+- added a dedicated release-notes draft for `v0.0.16` under `docs/release-notes-0.0.16.md`
+
+### Changed
+
+- user-facing terminology now consistently uses **TUI** instead of legacy `fullscreen` wording, including CLI/help/docs text and internal module/type naming
+- markdown code blocks and bash tool previews in the TUI now render as raw fenced blocks for easier copying and pasting
+- tool availability, advertised schemas, transcript validation, and tool execution now follow a shared architecture with a single CLI `ToolRegistry`, centralized provider transcript repair, explicit tool lifecycle phases, lifecycle hook events, and mutation-aware scheduling for real tool execution
+- release builds and npm installs now prefer stripped/compressed native binaries, reducing download size and improving install/update behavior
+
+### Fixed
+
+- `browser_fetch` now gives actionable missing-browser diagnostics and setup guidance instead of failing opaquely when no local browser is installed
+- TUI footer context usage no longer gets stuck at misleading values like `0.0%/272k (auto)` after resume/rebuild/fork paths
+- queued prompt previews no longer hide compaction/local-action status, and fast local actions no longer flash meaningless `0ms` / `0.0s` timing
+- failed or interrupted tool calls no longer wedge later prompts with provider-side `No tool output found ...` errors; transcript repair and lifecycle finalization now guarantee a valid follow-up turn state
+- join-timeout warnings no longer make an active TUI turn look finished before the runner actually completes
+- clean `bb-cli` builds no longer emit dead-code warnings from the new tool registry surface, and secret-bearing auth/token structs now redact sensitive fields in debug output instead of exposing raw credentials
+
+### Improved
+
+- ongoing Rust audit cleanup significantly reduced boolean-heavy APIs, overexposed state, and hotspot modules across `bb-cli`, `bb-core`, `bb-tools`, `bb-session`, `bb-tui`, `bb-provider`, `bb-plugin-host`, and `bb-hooks`, with expanded focused regression coverage throughout
+- npm install now prefers compressed `.gz` GitHub release assets when available, falling back to the legacy uncompressed binaries for older releases
+- release builds now strip debug info before publishing, substantially reducing native binary download size for npm installs and updates
+
+## [0.0.15] - 2026-04-12
+
+### Fixed
+
+- tui bash output now streams live stdout/stderr and keeps running, finished, historical, and expanded/collapsed tool blocks visually consistent with width-aware tail previews and elapsed/took footer hints
+- bash tool titles now skip shell prelude lines like `set -e`, show configured timeout values in the UI, and validate invalid timeout arguments instead of accepting zero or non-finite values
+- fixed the session getting stuck after interrupted or failed tool calls by flushing synthetic tool results before later prompts, preventing follow-up turns from failing with missing tool output
+- Codex tool-call request handling is more robust: tool calls are serialized sequentially, orphan tool results are sanitized out of requests, and streamed/done function-call events are deduplicated more safely
+- plain URLs and hyphenated text no longer trigger accidental markdown horizontal-rule rendering while explicit markdown rules and setext headings still render correctly
+- tui resume menu handling now awaits the async resume path correctly instead of dropping back through the synchronous menu flow
+
+### Added
+
+- regression coverage for interrupted tool-call recovery, Codex orphan-tool sanitization, builtin tool-name normalization, tui bash rendering consistency, and bash timeout validation/visibility
+
 ## [0.0.14] - 2026-04-12
 
 ### Added
 
-- fullscreen now supports extension-driven workflows and structured slash-command outcomes, including menus, hidden dispatches, and richer command result handling
-- `/settings` in fullscreen now exposes compaction controls for `Auto-compact`, `Reserve tokens`, and `Keep recent tokens`
+- tui now supports extension-driven workflows and structured slash-command outcomes, including menus, hidden dispatches, and richer command result handling
+- `/settings` in tui now exposes compaction controls for `Auto-compact`, `Reserve tokens`, and `Keep recent tokens`
 - skills can now be listed, disabled, and re-enabled from the CLI without deleting their installed files
 - startup model selection now prefers configured provider/model defaults more consistently, with better OpenAI startup fallback behavior
 - added a parity test script against installed pi compaction logic to keep BB token accounting aligned with upstream behavior
@@ -18,39 +64,39 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Fixed
 
 - session resume now restores the prior model and thinking level instead of starting with mismatched runtime defaults
-- fullscreen/TUI terminal rendering now sanitizes terminal control text more reliably and avoids ANSI leakage into the UI
+- tui/TUI terminal rendering now sanitizes terminal control text more reliably and avoids ANSI leakage into the UI
 - auto-compaction token estimation now matches pi more closely by using the last successful assistant usage plus trailing estimates, using ceil-based token heuristics, computing `tokens_before` from rebuilt context instead of raw payload size, and ignoring assistant usage from before the latest compaction boundary
-- fullscreen compaction behavior and status reporting are more consistent after auto-compaction and manual compaction events, and local fullscreen actions now show an animated elapsed-time status while they run
+- tui compaction behavior and status reporting are more consistent after auto-compaction and manual compaction events, and local tui actions now show an animated elapsed-time status while they run
 
 ### Changed
 
-- fullscreen extension workflows and session compaction support are now merged into the main interaction path on `master`
+- tui extension workflows and session compaction support are now merged into the main interaction path on `master`
 
 ## [0.0.13] - 2026-04-09
 
 ### Added
 
-- fullscreen screenshot and image clipboard paste now works on the normal paste path, with macOS clipboard fallbacks and Codex image preservation so pasted images reach image-capable models correctly
+- tui screenshot and image clipboard paste now works on the normal paste path, with macOS clipboard fallbacks and Codex image preservation so pasted images reach image-capable models correctly
 - model registry metadata now tracks image input capability, making `/models` truthful about image support and allowing runtime warnings when users attach images to text-only models
 
 ### Fixed
 
-- fullscreen clipboard image attach no longer leaks helper `true` / `false` output or stray follow-up paste text into the input block
+- tui clipboard image attach no longer leaks helper `true` / `false` output or stray follow-up paste text into the input block
 - attached image chips can now be removed with `Backspace`, image-only prompts can be submitted, and optimistic user messages keep attachment chip previews in the transcript
-- rebuilt fullscreen session transcripts now preserve user image attachment markers instead of silently dropping image blocks
+- rebuilt tui session transcripts now preserve user image attachment markers instead of silently dropping image blocks
 - managed `bb-clipboard-*.png` temp files are now cleaned up after removal or ingestion instead of lingering in `/tmp`
-- the fullscreen input block now hides raw `@file` tokens when the corresponding attachment chip is already shown, preventing duplicated `@file` text in the editor
-- fullscreen tool-header regression tests now match the intended live bash-header rendering and running-dot animation behavior
+- the tui input block now hides raw `@file` tokens when the corresponding attachment chip is already shown, preventing duplicated `@file` text in the editor
+- tui tool-header regression tests now match the intended live bash-header rendering and running-dot animation behavior
 
 ## [0.0.12] - 2026-04-06
 
 ### Fixed
 
-- direct `@image` references in print mode and fullscreen now attach real image inputs instead of falling back to UTF-8 read warnings
+- direct `@image` references in print mode and tui now attach real image inputs instead of falling back to UTF-8 read warnings
 - `@path with spaces` parsing now correctly keeps the full file path before trailing prompt text, including whole-message forms
 - image tool results are now preserved through provider conversion so models can actually see images returned by tools instead of responding as if no image was provided
-- fullscreen `@` folder navigation now keeps the completion menu open when you select a directory and immediately shows the next level, including directories with spaces
-- the fullscreen input block now shows attached files as `[name, sizeKB]`, keeps those chips visible, and places the cursor below them so typing starts after the attachments
+- tui `@` folder navigation now keeps the completion menu open when you select a directory and immediately shows the next level, including directories with spaces
+- the tui input block now shows attached files as `[name, sizeKB]`, keeps those chips visible, and places the cursor below them so typing starts after the attachments
 
 ### Changed
 
@@ -60,9 +106,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
-- startup update notices in the fullscreen transcript are now highlighted so available updates stand out clearly during startup
-- read-tool line ranges in fullscreen tool activity now highlight the requested span, so values like `2148-2267/5006` stand out while the model is using tools
-- fullscreen footer and `/session` info now show the active execution posture so safety vs yolo is visible during a run
+- startup update notices in the tui transcript are now highlighted so available updates stand out clearly during startup
+- read-tool line ranges in tui tool activity now highlight the requested span, so values like `2148-2267/5006` stand out while the model is using tools
+- tui footer and `/session` info now show the active execution posture so safety vs yolo is visible during a run
 
 ### Improved
 
@@ -81,7 +127,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - npm install now uses a longer timeout, retries release-binary downloads, and reports real download errors instead of incorrectly saying no matching prebuilt binary exists
 - npm install now shows progress logs during native binary download and verification so first-time installs on macOS/Linux are less confusing
-- fullscreen `/login` provider-family status now correctly shows OpenAI OAuth state after ChatGPT login instead of incorrectly showing the API key path as not authenticated
+- tui `/login` provider-family status now correctly shows OpenAI OAuth state after ChatGPT login instead of incorrectly showing the API key path as not authenticated
 
 ### Changed
 
@@ -97,9 +143,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
-- fullscreen paste in iTerm2/SSH no longer corrupts the input area after paste
+- tui paste in iTerm2/SSH no longer corrupts the input area after paste
 - pasted file and image paths are normalized more reliably, including quoted paths and `file://` URLs
-- fullscreen prompt submission now expands `@file` references consistently
+- tui prompt submission now expands `@file` references consistently
 - running tool timers continue updating after `TurnEnd` while tools are still executing
 - sub-second tool durations now display as `ms` instead of `0.0s`
 - startup Skills/Prompts/Extensions note now only appears at startup or explicit `/reload`
@@ -107,7 +153,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Changed
 
-- fullscreen `Ctrl+V` now falls back to clipboard text when no clipboard image is available
+- tui `Ctrl+V` now falls back to clipboard text when no clipboard image is available
 - `@` autocomplete now inserts quoted file references when paths contain spaces
 
 ## [0.0.8] - 2026-04-06
@@ -128,6 +174,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - latest published package includes the post-0.0.7 startup, auth, model-default, and update-notice improvements
 
+[0.0.16]: https://github.com/shuyhere/bb-agent/releases/tag/v0.0.16
+[0.0.15]: https://github.com/shuyhere/bb-agent/releases/tag/v0.0.15
 [0.0.14]: https://github.com/shuyhere/bb-agent/releases/tag/v0.0.14
 [0.0.13]: https://github.com/shuyhere/bb-agent/releases/tag/v0.0.13
 [0.0.12]: https://github.com/shuyhere/bb-agent/releases/tag/v0.0.12

package/README.md

@@ -4,7 +4,7 @@
 
 > BB means Bridge Baby in Death Stranding. I named this project that way because while building it, I was also enjoying Death Stranding and loved the idea of connecting everyone together.
 
-A Rust-native AI coding agent for the terminal — featuring a fullscreen TUI, multi-provider support, tool use, session persistence, branching, extensions, and skills.
+A Rust-native AI coding agent for the terminal — featuring a TUI, multi-provider support, tool use, session persistence, branching, extensions, and skills.
 
 ## Install
 
@@ -17,7 +17,7 @@ npm install -g @shuyhere/bb-agent
 npm install downloads a small wrapper package first, then fetches the matching native BB-Agent binary from the GitHub release for your platform.
 
 What to expect:
-- first install can take a bit because npm downloads and verifies the native binary
+- first install can take a bit because npm downloads a compressed native binary, expands it locally, and verifies it
 - the installer now prints progress and retry information while downloading
 - after install, run `bb`
 
@@ -68,7 +68,7 @@ Inside the TUI, run:
 /login
 ```
 
-This opens the provider picker and auth flow directly in the fullscreen UI.
+This opens the provider picker and auth flow directly in the TUI.
 
 If you prefer, you can also log in from a normal terminal:
 
@@ -81,12 +81,12 @@ bb login google       # Login to Google (API key)
 
 Or set environment variables: `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `GOOGLE_API_KEY`, etc.
 
-That's it! Run `bb` to launch the fullscreen interactive terminal UI. Type your prompt and press Enter.
+That's it! Run `bb` to launch the interactive TUI. Type your prompt and press Enter.
 
 ### More ways to use `bb`
 
 ```bash
-bb                                    # Launch the fullscreen TUI
+bb                                    # Launch the TUI
 bb "Explain this codebase"            # TUI with an initial prompt
 bb -p "What is 2+2?"                  # Print mode (non-interactive, pipe-friendly)
 bb -c                                 # Continue your last session
@@ -98,7 +98,7 @@ bb --list-models                      # List all available models
 
 ## Features
 
-- **Fullscreen TUI** — rich terminal interface with streaming output, markdown rendering, syntax highlighting
+- **TUI** — rich terminal interface with streaming output, markdown rendering, syntax highlighting
 - **Multi-provider** — Anthropic (Claude), OpenAI, Google (Gemini), Groq, xAI, OpenRouter, and custom OpenAI-compatible endpoints
 - **Built-in tools** — `read`, `write`, `edit`, `bash`, `find`, `grep`, `ls`, `web_search`, `web_fetch`, `browser_fetch`
 - **Safety and yolo execution modes** — default safety posture restricts built-in `write` and `edit` to the active workspace; yolo removes that guard
@@ -165,7 +165,7 @@ BB-Agent uses layered configuration:
 
 ### Execution Modes
 
-BB-Agent exposes the active permission posture in fullscreen and `/session`.
+BB-Agent exposes the active permission posture in the TUI and `/session`.
 
 - `safety` is the default. Built-in `write` and `edit` stay inside the current workspace, and bash commands use the safer approval/sandboxed posture.
 - `yolo` is the opt-in less-restrictive mode.
@@ -202,14 +202,14 @@ Example:
 | `bb-provider` | Model/provider integrations and streaming |
 | `bb-hooks` | Hook event types for extensions |
 | `bb-plugin-host` | Plugin discovery and host runtime |
-| `bb-tui` | Terminal UI components and fullscreen experience |
+| `bb-tui` | Terminal UI components and the interactive TUI experience |
 | `bb-cli` | The `bb` command-line application |
 
 ## Troubleshooting
 
 ### Terminal & Font Compatibility
 
-BB-Agent uses Unicode glyphs and ANSI color in the fullscreen TUI. For the best visual experience, use a modern terminal and a Unicode-capable monospace font such as:
+BB-Agent uses Unicode glyphs and ANSI color in the TUI. For the best visual experience, use a modern terminal and a Unicode-capable monospace font such as:
 
 - JetBrains Mono
 - SF Mono / Menlo

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@shuyhere/bb-agent",
-  "version": "0.0.14",
+  "version": "0.0.16",
   "description": "BB-Agent — a Rust-native AI coding agent for the terminal",
   "license": "MIT",
   "repository": {

package/scripts/postinstall.js

@@ -8,6 +8,7 @@ const path = require("path");
 const os = require("os");
 const https = require("https");
 const http = require("http");
+const zlib = require("zlib");
 
 const packageJson = require("../package.json");
 const BINARY_RELEASE_TAG = `v${packageJson.version}`;
@@ -51,6 +52,20 @@ function assetNameForTarget(target) {
   return isWindows() ? `bb-${target}.exe` : `bb-${target}`;
 }
 
+function assetCandidatesForTarget(target) {
+  const assetName = assetNameForTarget(target);
+  return [
+    {
+      assetName: `${assetName}.gz`,
+      compressed: true,
+    },
+    {
+      assetName,
+      compressed: false,
+    },
+  ];
+}
+
 function logLine(message = "") {
   try {
     process.stderr.write(`${message}\n`);
@@ -387,9 +402,16 @@ function verifyBinary(binaryPath) {
   return { ok: true, version };
 }
 
+function expandCompressedBinary(src, dest) {
+  logLine("Decompressing downloaded BB-Agent binary...");
+  ensureParentDir(dest);
+  const expanded = zlib.gunzipSync(fs.readFileSync(src));
+  fs.writeFileSync(dest, expanded);
+  removeIfExists(src);
+}
+
 async function tryDownloadPrebuilt(target) {
-  const assetName = assetNameForTarget(target);
-  const url = `https://github.com/${REPO}/releases/download/${BINARY_RELEASE_TAG}/${assetName}`;
+  const assetCandidates = assetCandidatesForTarget(target);
 
   fs.mkdirSync(NATIVE_DIR, { recursive: true });
   const dest = nativeBinaryPath();
@@ -412,37 +434,67 @@ async function tryDownloadPrebuilt(target) {
         `Downloading BB-Agent ${BINARY_RELEASE_TAG} for ${target} (attempt ${attempt}/${MAX_DOWNLOAD_ATTEMPTS})...`
       );
       logLine("This may take a little while on first install because npm downloads the native binary from the GitHub release.");
-      removeIfExists(tmpDest);
-      await requestBinary(url, tmpDest, 0);
-      if (!isWindows()) {
-        fs.chmodSync(tmpDest, 0o755);
-      }
 
-      const verified = verifyBinary(tmpDest);
-      if (!verified.ok) {
+      let missingCompressedAsset = false;
+      for (const asset of assetCandidates) {
+        const url = `https://github.com/${REPO}/releases/download/${BINARY_RELEASE_TAG}/${asset.assetName}`;
+        const downloadDest = asset.compressed ? `${tmpDest}.gz` : tmpDest;
+
         removeIfExists(tmpDest);
-        return {
-          ok: false,
-          kind: "verify",
-          message: `Downloaded binary could not run: ${verified.message}`,
-        };
-      }
+        removeIfExists(`${tmpDest}.gz`);
 
-      fs.renameSync(tmpDest, dest);
-      refreshCacheFromExistingBinary(target, dest);
-      logLine("Cached verified BB-Agent binary for future installs.");
-      logLine("✓ BB-Agent binary installed successfully.");
-      return { ok: true, source: "download" };
+        try {
+          if (asset.compressed) {
+            logLine(`Trying compressed release asset ${asset.assetName} first for a faster download.`);
+          }
+          await requestBinary(url, downloadDest, 0);
+          if (asset.compressed) {
+            expandCompressedBinary(downloadDest, tmpDest);
+          }
+          if (!isWindows()) {
+            fs.chmodSync(tmpDest, 0o755);
+          }
+
+          const verified = verifyBinary(tmpDest);
+          if (!verified.ok) {
+            removeIfExists(tmpDest);
+            return {
+              ok: false,
+              kind: "verify",
+              message: `Downloaded binary could not run: ${verified.message}`,
+            };
+          }
+
+          fs.renameSync(tmpDest, dest);
+          refreshCacheFromExistingBinary(target, dest);
+          logLine("Cached verified BB-Agent binary for future installs.");
+          logLine("✓ BB-Agent binary installed successfully.");
+          return { ok: true, source: "download" };
+        } catch (err) {
+          lastError = err;
+          removeIfExists(tmpDest);
+          removeIfExists(`${tmpDest}.gz`);
+          if (err.kind === "not-found") {
+            if (asset.compressed) {
+              missingCompressedAsset = true;
+              logLine(`Compressed asset ${asset.assetName} not found; falling back to the uncompressed release binary.`);
+              continue;
+            }
+            return {
+              ok: false,
+              kind: "not-found",
+              message: missingCompressedAsset
+                ? `No release asset named ${asset.assetName} or ${assetCandidates[0].assetName} was found for ${BINARY_RELEASE_TAG}.`
+                : `No release asset named ${asset.assetName} was found for ${BINARY_RELEASE_TAG}.`,
+            };
+          }
+          throw err;
+        }
+      }
     } catch (err) {
       lastError = err;
       removeIfExists(tmpDest);
-      if (err.kind === "not-found") {
-        return {
-          ok: false,
-          kind: "not-found",
-          message: `No release asset named ${assetName} was found for ${BINARY_RELEASE_TAG}.`,
-        };
-      }
+      removeIfExists(`${tmpDest}.gz`);
       if (attempt < MAX_DOWNLOAD_ATTEMPTS) {
         logLine(`Download failed (${err.message}). Retrying...`);
         await new Promise((resolve) => setTimeout(resolve, 1000 * attempt));