Taber v0.6.2

English

Fixed

• Agent actions tied to the current tab now run in order, and task event persistence failures stop the task before duplicate page actions can occur.
• Exported documents no longer load remote resources embedded in workspace HTML.
• Event timelines retain their causal order even when system clocks move backward.

Changed

• Side-panel workspace files refresh only when a relevant file mutation occurs; runtime acceptance now verifies a live file write appears without reloading the panel.
• Event projection is linear for large session histories, reducing side-panel work as logs grow.
• The full runtime gate now covers the extension, Browser REPL, and side panel in an isolated build.
• Added regression coverage for agent instructions, context compaction, and xAI subscription runtime behavior.

中文

修复

• 与当前标签页绑定的 Agent 操作现在按顺序执行；任务事件持久化失败会在重复页面操作前终止任务。
• 导出的文档不再加载工作区 HTML 中嵌入的远程资源。
• 即使系统时钟回拨，事件时间线也会保持因果顺序。

变更

• 侧边栏工作区文件只在相关文件变更时刷新；运行时验收会验证实时写入无需重新加载侧栏即可显示。
• 大型会话历史的事件投影改为线性处理，减少日志增长时的侧栏开销。
• 完整运行时门禁现在在隔离构建中覆盖扩展、Browser REPL 和侧边栏。
• 新增 Agent 指令、上下文压缩和 xAI 订阅运行时行为的回归覆盖。

Taber v0.6.1

English

Fixed

• Browser actions recover from malformed refs when the model also provides one unambiguous semantic locator, preventing repeated INVALID_TARGET loops on dynamic iframe pages.

Changed

• Browser REPL guidance documents the { count, matches, candidates } result returned by queryText("text").

中文

修复

• 模型同时提供唯一可靠的语义定位器时，即使 ref 被污染，浏览器操作也会自动使用语义定位器，避免动态 iframe 页面反复进入 INVALID_TARGET 循环。

变更

• Browser REPL 文档补充 queryText("text") 返回的 { count, matches, candidates } 结构。

Taber v0.6.0

English

Highlights

• Added an optional personal profile for form filling. The profile stays local, is exposed only as a read-only /profile.md file, and requires explicit consent for each task.
• Side-panel history now restores the conversation selected by the user, including a blank new conversation, after the panel is closed and reopened.
• The side-panel shortcut now uses Chrome's native action toggle, so one browser-managed shortcut opens and closes the panel without extension-side state races.

Changed

• The shortcut pill reads the active binding from chrome.commands.getAll() and refreshes when the side panel regains focus.
• Browser REPL guidance and runtime results now use serializable element descriptors, exact selectors, and clearer errors for invented refs, bare indexes, and helper misuse.
• Browser form matching recognizes labels from single-field component containers while avoiding ambiguous multi-field rows.

Fixed

• Session selection writes and snapshot reads are serialized and transactionally coordinated, preventing stale events or delayed task startup from overwriting the user's last selection.
• Personal profile reads are redacted in persisted tool events, tool runs, evidence digests, timelines, and exports while the authorized model call still receives the content.

中文

重点

• 新增可选个人资料，用于填表。资料只保存在本地，仅以只读 /profile.md 暴露，并且每个任务都必须显式授权。
• 侧边栏关闭再打开时恢复用户最后选择的会话，包括空白新会话。
• 侧边栏快捷键改用 Chrome 原生 action 切换，同一个浏览器快捷键可打开和关闭侧栏，不再依赖扩展侧状态猜测。

变更

• 快捷键胶囊从 chrome.commands.getAll() 读取当前绑定，侧栏重新获得焦点后自动刷新。
• Browser REPL 统一使用可序列化元素描述符、精确选择器，并对伪造 ref、裸索引和 helper 误用返回明确错误。
• 表单匹配支持从单字段组件容器读取可见标签，同时避免把多字段行误判为同一字段。

修复

• 会话选择写入与快照读取串行并在事务中协调，避免旧事件或延迟任务启动覆盖用户最后选择。
• 个人资料读取在持久化工具事件、工具运行记录、证据摘要、时间线和导出中统一脱敏；已授权的当前模型调用仍可读取原文。

Taber v0.5.2

English

Changed

• Localized the extension description for English and Simplified Chinese browser interfaces.

中文

变更

• 扩展描述现已适配英文和简体中文浏览器界面。

Taber v0.5.1

English

Changed

• Canonicalized model tool inputs so task targets, browser execution timeouts, and runtime-only fields remain host-owned; irrelevant placeholder fields are discarded before execution.
• Hardened Browser REPL lifecycle management with one total budget, cancellation propagation to in-flight page commands, remaining-budget caps for helpers, and explicit NO_EVIDENCE results.
• Navigation failures now return recoverable error contracts with retry guidance, while side-panel presentation distinguishes recoverable failures from task failures.

Diagnostics

• Added runtime.configured session diagnostics for non-secret provider/model metadata, reasoning effort, and tool schema version.

Documentation

• Documented canonical tool input boundaries and the limited Chrome offscreen runtime API.

中文

变更

• 规范化模型工具输入：目标标签页、浏览器执行超时和运行时字段由宿主持有，模型填入的无关占位字段会在执行前丢弃。
• 加固 Browser REPL 生命周期：统一总预算、取消传播到在途页面命令、按剩余预算限制 helper，并在没有证据时明确返回 NO_EVIDENCE。
• 导航失败改为带重试建议的可恢复错误契约；侧边栏区分可恢复工具失败与任务失败。

诊断

• 新增 runtime.configured 会话诊断事件，记录非秘密的 provider/model 元数据、推理强度和工具 Schema 版本。

文档

• 补充模型工具输入边界及 Chrome offscreen 受限 runtime API 说明。

Taber v0.5.0

English

Highlights

• Added an optional Follow AI actions mode. It activates the controlled target tab before page operations so users can watch the Agent work, without bringing the Chrome window to the front. Background mode remains the default.
• The selected mode is persisted locally and captured when a task starts, so a running task keeps one consistent execution policy.

Changed

• Foreground mode applies consistently to page reading, image extraction, browser control, Browser REPL, debug tools, and navigation. navigate.open target:"new" and navigate.switchTab follow the task mode.
• Background viewport capture still activates the target only when Chrome requires it, then restores the previous tab when the user has not switched tabs during capture.
• The obsolete active navigation input is rejected; tab activation is controlled by the task mode instead of model-generated input.

Security

• Runtime side-panel smoke tests verify the isolated build marker before clearing extension data or changing site permissions, preventing a test run from modifying an unrelated installed Taber.

中文

重点

• 新增可选的“跟随 AI 操作”模式。页面工具执行前会激活受控目标标签页，方便用户观看 Agent 操作，但不会把 Chrome 窗口抢到操作系统前台；默认仍为后台模式。
• 模式偏好保存在本地，并在任务启动时固定，运行中的任务始终使用同一执行策略。

变更

• 前台模式统一应用于页面读取、图片提取、浏览器控制、Browser REPL、调试工具和导航；navigate.open target:"new" 与 navigate.switchTab 也遵循任务模式。
• 后台 viewport 截图仍会在 Chrome 必须激活标签页时临时激活目标；如果用户期间没有切换标签页，截图后会恢复原标签页。
• 废弃的 active 导航输入会被拒绝；标签激活由任务模式控制，不再由模型输入决定。

安全

• 侧边栏运行时烟测会先校验本次隔离构建标记，再清空扩展数据或修改站点权限，避免测试误操作用户已安装的其他 Taber。

Taber v0.4.1

English

Fixed

• Browser Control centers the permission guide after the settings dialog enters, keeps the full spotlight visible, and returns to the top after authorization. Reduced-motion mode positions it immediately.
• Activity group labels and colors follow the latest or terminal outcome. Recovered and completed groups use the success tone; failed steps remain visible in details.

Changed

• The Sources card uses one edge-to-edge hover surface with reversible motion and a subtle icon lift. Its popover opens directly into source items without a repeated heading.
• Browser ref guidance, schemas, and runtime tests match the implemented lifecycle: refs survive snapshots and ordinary DOM updates while the same target stays visible and unchanged.

Performance

• AgentHost loads browser tools and provider-specific model runtimes when a task needs them, reducing the initial offscreen entry from 629 kB to 361 kB.
• Release verification enforces a 500 kB initial-entry budget for AgentHost and the side panel.

Documentation

• README adds lossless product screenshots for the idle side panel, completed tasks, and site skills.

中文

修复

• 浏览器控制在设置弹窗进场后将权限引导完整移至可视区域中央，授权完成后回到顶部。减少动态效果模式会立即定位。
• 活动组文案和颜色跟随最新步骤或整轮终态。恢复成功和已完成状态使用成功色，失败步骤仍保留在详情中。

变更

• 来源卡片使用覆盖整张卡片的单一 hover 表面，并统一进出过渡与图标微动效。气泡直接展示来源条目，不再重复标题。
• 浏览器 ref 文案、Schema 和运行时测试与实际生命周期对齐：同一目标保持可见且未变化时，ref 可跨快照和普通 DOM 更新继续使用。

性能

• AgentHost 仅在任务需要时加载浏览器工具和对应供应商模型运行时，offscreen 初始入口由 629 kB 降至 361 kB。
• 发布校验对 AgentHost 和侧边栏初始入口执行 500 kB 体积预算。

文档

• README 增加空闲侧边栏、任务结果和站点技能的无损产品截图。

Taber v0.4.0

English

Highlights

• Rebuilt the side-panel timeline around compact activity groups that combine reasoning and tool work without hiding execution evidence.
• Hardened long-running Agent tasks with bounded steps, per-step and total timeouts, stream health checks, and explicit terminal failures.
• Expanded builtin site skills from 8 to 27, organized into ticketing, shopping, social, video, travel, developer, and reference categories.
• Added GPT-5.6 Sol, Terra, and Luna support across OpenAI-compatible and ChatGPT/Codex flows.
• Decoupled task startup from page access: pure Q&A can begin on new-tab or restricted pages, then navigate to an operable page.
• Added exportable session diagnostics so persisted Agent behavior can be inspected without reproducing the task.

Added

• Activity groups with running, completed, failed, stopped, and recoverable-warning states; each group reports step count and elapsed duration.
• Expandable activity details for every reasoning and tool step, including copyable raw tool input, output, and error evidence.
• Activity auto-follow: opening a group jumps to the newest step, follows live additions, pauses when the user scrolls up, and resumes near the bottom.
• One-line JSONL session export from Settings → Preferences → Diagnostics. Strings over 2,048 characters and screenshot data URLs are truncated to reviewable previews.
• scripts/dump-session-events.ts for listing and exporting sessions from a CDP-reachable Taber instance, plus docs/debugging.md with event signatures and failure diagnosis.
• Stream guards for malformed tool JSON, content after complete JSON, mismatched delimiters, oversized tool input, oversized step output, and degenerate whitespace runs.
• A 180-second stream idle timeout, 5-minute per-step timeout, 30-minute task timeout, and explicit 20-step tool-loop ceiling.
• Seven localized skill categories with per-skill and per-category enable controls; custom user/Agent skills remain a separate first-class group.
• Nineteen new builtin skills for Tixbay, Damai, Ticketmaster, StubHub, Amazon, Taobao, JD, eBay, X, Xiaohongshu, Weibo, Zhihu, Instagram, YouTube, Bilibili, Douyin, TikTok, Booking.com, and Airbnb.
• Chinese names and descriptions for all 27 builtin skills, while stored skill identifiers remain stable.
• Prompt-based skill pre-matching: tasks can discover a site skill from a site name in the request before the first navigation.
• GPT-5.6 Sol, Terra, and Luna presets with 1,050,000-token context windows; OpenRouter also includes openai/gpt-5.6-sol.
• Weekly staleness checks for the cached models.dev catalog, refreshed in the background from provider settings.
• ADR 0018 documenting task startup and recovery from non-operable target tabs.

Changed

• Consecutive reasoning and tool events now render as one activity block; assistant text splits blocks and keeps the original event order.
• Cancelled tasks render as stopped rather than looking active; recoverable tool results render as warnings instead of false success or fatal failure.
• Live side-panel events insert by event ID and deduplicate; session snapshots also read Agent events by ID rather than wall-clock timestamps.
• Quick actions use a stable Summarize, Skills, Research, Compare order instead of guessing intent from the current URL or page title.
• The Skills panel groups builtins by category, exposes localized metadata, supports whole-group toggles, and keeps custom skills visible for discovery.
• Matching skills are mandatory prior reading. Tasks that need at least four exploratory tool calls to discover a reusable route must save or update a skill before finishing.
• Builtin skills seed before the side panel becomes ready, so the Skills panel is populated before the first Agent task creates an offscreen host.
• Tasks lock any active tab that exists, including chrome://, edge://, and new-tab pages; page-access failures remain recoverable while missing or closed targets remain fatal.
• Page tools operate on the locked target in the background without activating it on every call.
• navigate.switchTab changes the task target without stealing focus; navigate.open target:"new" opens in the background unless active:true is requested.
• Viewport screenshots temporarily activate the target because Chrome only captures the visible tab, then restore the prior tab only if the user did not switch during capture.
• Navigation load timeouts return navigation.status:"timeout" with inspectable tab state. Superseded net::ERR_ABORTED navigations keep waiting for the replacement navigation.
• Browser refs survive the immediately preceding routed snapshot and ordinary page mutations while marker, visibility, frame URL, tag, and fingerprint checks still reject changed targets.
• Structured browser target parsing drops empty placeholder fields, prefers an explicit snapshot ref over echoed semantic fields, and lets semantic locators outrank placeholder coordinates.
• getDocument and extractImage ignore irrelevant padded fields from model-generated tool input while preserving required-field and value validation.
• Model lists sort newer numeric generations first, then provider priority; subscription model chips can expand from the preview to the full list.
• Codex OAuth, model discovery, and inference identify as official codex_cli_rs client 0.144.1, including the inference version header required by gated models.
• OpenAI and Codex requests ask for detailed reasoning summaries only on models that declare reasoning support; non-reasoning models receive no reasoning fields.
• Side-panel icons now use the existing Lucide package throughout. The unused phosphor-svelte production dependency was removed.
• Settings, provider onboarding, session history, composer, sources, files, toasts, and subscription cards share one motion and interaction system with reduced-motion fallback.
• Session history positions the current item below its sticky action row and uses bidirectional view transitions for old, new, and rapidly switched sessions.

Fixed

• Streaming text, reasoning, and tool-input deltas flush at 512 characters or within 300 ms, reducing IndexedDB transactions without waiting for another delta.
• Buffered event types and IDs flush in sequence before lifecycle events, preserving the exact persisted and live event order.
• Out-of-order runtime broadcasts no longer disappear, and duplicate broadcasts no longer create duplicate timeline parts.
• Invalid model tool input now creates a persisted tool.failed event instead of leaving the activity permanently pending.
• Active partial tool inputs receive failure events when a stream health guard stops generation.
• A stream ending at the tool-loop limit now fails with a narrow-task retry message instead of reporting an empty or successful response.
• Internal timeout aborts are distinguished from user cancellation and reported as execution timeouts.
• Provider errors retain up to 600 characters of upstream response context after credential redaction, preserving messages such as “Model not found.”
• Remote document fetches stop after 30 seconds and retain HTTP, CORS, timeout, or network reasons in recoverable results.
• Page-access and screenshot failures retain the browser’s original reason alongside Browser Control guidance.
• Explicit mismatched tabId values fail before browser access in getDocument, all extractImage sources including viewport, browser, browserRepl, debug-only debugger, and every non-switch navigation action.
• The background broker and navigate controller enforce the same target boundary, so direct internal calls cannot bypass Agent runtime validation.
• A user tab switch before or during viewport capture rejects the screenshot and never restores focus over the user’s newer selection.
• Non-operable targets no longer trigger fatal task cancellation; only a closed or missing target ends the task, with no fallback to another active tab.
• Navigation no longer activates the target for open, back, forward, reload, current-tab reads, debugger calls, or page-tool validation.
• fs read trims accidental model output after a .md skill path instead of turning a valid skill lookup into a not-found error.
• Session switching rejects stale async reads, so rapid History → New Session actions cannot be overwritten by an older request.
• Stopped and completed activity groups no longer retain running beam or breathing effects.
• Skill reads and writes use a distinct “site skill” label instead of appearing as generic file operations.
• Side-panel runtime smoke now handles Chrome 150 API namespaces, CSS-transformed labels, icon-only controls, and short history lists without weakening user-facing assertions.

Security

• Only navigate.switchTab, navigate.open target:"new", or explicit side-panel confirmation may retarget a task; all other cross-target IDs fail with a recovery instruction.
• Starting from a restricted page does not widen host permissions or page access. Page tools fail explicitly until navigation reaches an allowed page.
• Release builds still omit debugger and cookies permissions. Detailed reasoning uses provider summaries and does not expose raw chain-of-thought.

中文

重点

• 侧边栏时间线改为紧凑活动组，把 reasoning 与工具过程收进同一条执行轨迹，同时保留完整证据。
• Agent 长任务新增步数、单步/总时长、流健康检查和明确终止错误，不再无限生成或假完成。
• 内置站点技能从 8 个扩展到 27 个，按票务、购物、社交、视频、旅行、开发和知识七类管理。
• OpenAI-compatible 与 ChatGPT/Codex 流程新增 GPT-5.6 Sol、Terra、Luna 支持。
• 任务启动与页面访问解耦：可从新标签页或受限页面开始纯问答，再导航到可操作页面。
• 新增可导出的会话诊断，不必复现任务即可检查已持久化的 Agent 行为。

新增

• 活动组支持运行、完成、失败、停止和可恢复警告五种状态，显示步骤数与执行时长。
• 每个 reasoning 和工具步骤均可展开，保留可复制的原始工具输入、输出与错误证据。
• 活动自动跟随：展开后跳到最新步骤；实时追加时持续跟随；用户上翻后暂停，回到底部恢复。
• 设置 → 偏好 → 诊断新增 JSONL 会话导出。超过 2,048 字符的字符串和截图 data URL 会截成可审阅预览。
• 新增 scripts/dump-session-events.ts，可从支持 CDP 的 Taber 实例列出并导出会话；docs/debugging.md 记录事件特征和故障判读。
• 流守护覆盖非法工具 JSON、完整 JSON 后续写、括号不匹配、超大工具入参、超大单步输出和退化空白输出。
• 新增 180 秒流空闲超时、5 分钟单步超时、30 分钟任务超时和明确的 20-step ...

Taber v0.3.0

English

Highlights

• Added the fs tool: a virtual file system exposing a per-session /workspace file area and the /skills site-skill library through one ls/read/write surface.
• Added the site skills system: reusable, agent-authored site knowledge with lazy disclosure, host-change announcements, builtin seeds for 8 major sites, and a freshness loop.
• Added fetch-first URL reading: getDocument source:"url" fetches public webpages and PDFs directly without opening a tab, making cross-page collection parallel.
• Added a document workflow: upload PDF/Word/text files, generate Markdown/HTML/CSV/Word from tasks, download from the side panel, and export Markdown/HTML to PDF via the browser print dialog.
• Added context economy: oversized document and REPL outputs are saved to /workspace and only a preview plus a stable file path enters the model context and event log.

Added

• fs tool (ls/read/write): /workspace/<name> holds session uploads and outputs; /skills/<slug>.md exposes site skills as frontmatter Markdown files, matching the file-system prior LLMs are trained on.
• File workspace: writing .md/.txt/.html/.csv/.json stores text; writing .docx converts Markdown to Word (docx, dynamically imported); uploaded .docx/.pdf/text files are readable as text through getDocument source:"file" (mammoth/pdfjs).
• Side panel file UI: paperclip upload in the composer (with attachment chips and prompt annotation), a files strip with download, delete, and Export PDF actions, and a dedicated print page (print.html, marked + DOMPurify) for high-quality PDF export via the browser print dialog.
• Site skills: task instructions carry a digest of skills matching the current site; navigate (all actions), browser, and browserRepl announce availableSkills whenever an action lands on a new host, and never repeat on the same host.
• Skill authoring loop: the agent saves non-obvious reusable site flows as /skills/*.md files with structured frontmatter (name/hosts/description); skills are priors and live page state always wins.
• Skill freshness loop: after two consecutive tool failures in a task that read skills, the error hints at revising the stale skill file (once per task).
• Builtin seed skills (versioned, idempotent): Hacker News, Reddit, GitHub, Wikipedia, Stack Exchange, npm registry, PyPI, and arXiv fetch-first API endpoints with data shapes, quota pitfalls, and fallbacks.
• Skills management UI in the side panel: list, enable/disable, and delete stored skills.
• getDocument source:"url": direct fetch of http/https pages and PDFs (content-type routed) reusing the Readability/Markdown/table pipeline; failures explicitly point back to navigate + currentPage.
• browser coordinate targets { x, y } (viewport CSS px): visual fallback for canvas/visual-only UIs, resolving through open shadow roots and failing clearly on iframes; viewport screenshots now include width/height in CSS px as the coordinate reference.
• Context spill: getDocument content over 12k chars and browserRepl JSON values over the same threshold are written to /workspace/saved-<hash>.md|json with a 4k preview, savedTo, and a read-back hint; identical content maps to the same file.
• Local database: skills table (v2), per-session files table pruned with sessions (v3), and a v4 repair migration that renames legacy skills whose names collide on the /skills/<slug>.md path (newest keeps its name, nothing deleted).
• ADR 0015 (agent-authored site skills), ADR 0016 (fs tool and file workspace), ADR 0017 (fetch-first URL reading).

Changed

• Fixed top-level tool set is now 6: getDocument, extractImage, navigate, browser, browserRepl, fs (debug builds still add debugger).
• Agent instructions v5: goal-anchored efficient loop (define done → smallest evidence → shortest path → act once → verify), fetch-first URL policy, and file/skill strategy.
• getDocument source:"pdf" merged into source:"url" (auto-detected); source:"file" now reads workspace files by name instead of accepting inline fileText.
• Product narrative updated to a powerful, all-capable side panel agent across CONTEXT/PRODUCT/plan; stop/resume and tool-trail mechanics unchanged.
• Store compliance disclosure now covers locally stored session files and site skills.
• New dependencies: docx and mammoth, both pure JS and dynamically imported so they load only when converting documents.

Fixed

• Builtin skill seeding no longer overwrites user- or agent-authored skills with the same name, and individual seed failures no longer abort the rest.
• Skill names that collide on the /skills/<slug>.md path are rejected at save time, and renaming a skill file is atomic (save first, then delete the old entry).
• Runtime smoke script asserted a hardcoded, outdated instructionsVersion; it now references the shared constant.

中文

重点

• 新增 fs 工具：虚拟文件系统，用一套 ls/read/write 同时暴露每会话的 /workspace 文件区和 /skills 站点技能库。
• 新增站点技能系统：Agent 自行沉淀的可复用站点知识，支持懒加载披露、跨站提示、8 个大站内置种子和时效闭环。
• 新增 fetch-first URL 阅读：getDocument source:"url" 不开标签页直接拓取公开网页和 PDF，跨页收集可并行。
• 新增文档工作流：上传 PDF/Word/文本，任务产出 Markdown/HTML/CSV/Word，侧边栏下载，Markdown/HTML 可通过浏览器打印导出 PDF。
• 新增上下文经济：超大的文档和 REPL 输出自动落盘到 /workspace，模型上下文和事件日志只保留预览和稳定文件路径。

新增

• fs 工具（ls/read/write）：/workspace/<name> 存会话上传与产出；/skills/<slug>.md 以 frontmatter Markdown 文件形态暴露站点技能，利用模型对文件系统操作的预训练先验。
• 文件工作区：写 .md/.txt/.html/.csv/.json 存文本；写 .docx 自动把 Markdown 转 Word（docx 库，按需加载）；上传的 .docx/.pdf/文本通过 getDocument source:"file" 解析为文本（mammoth/pdfjs）。
• 侧边栏文件 UI：输入框回形针上传（附件 chip + 提示词自动标注）；文件条支持下载、删除、导出 PDF；新增打印页（print.html，marked + DOMPurify）走浏览器原生排版导出高质量 PDF。
• 站点技能：任务指令附当前站点匹配技能摘要；navigate（全部动作）、browser、browserRepl 落到新 host 时统一附 availableSkills 提示，同 host 不重复。
• 技能沉淀闭环：Agent 把非显而易见、可复用的站点流程写成 /skills/*.md（结构化 frontmatter：name/hosts/description）；技能是先验，与页面实时状态冲突时以页面为准。
• 技能时效闭环：任务内读过技能后若工具连续失败两次，错误信息提示修订可能过时的技能文件（每任务只提示一次）。
• 内置种子技能（按版本幂等写入）：Hacker News、Reddit、GitHub、Wikipedia、Stack Exchange、npm、PyPI、arXiv 的 fetch-first API 端点，含数据结构、配额陷阱和回退方案。
• 侧边栏技能管理 UI：列表、启用/停用、删除。
• getDocument source:"url"：直接 fetch http/https 网页与 PDF（按 content-type 自动分流），复用 Readability/Markdown/表格解析管线；失败时明确引导回退 navigate + currentPage。
• browser 坐标 target { x, y }（viewport CSS px）：canvas/纯视觉 UI 的视觉兜底，递归穿透 open shadow root，命中 iframe 时明确失败；视口截图结果附 CSS px 的 width/height 作为坐标基准。
• 上下文落盘：getDocument 内容超 12k 字符、browserRepl JSON 结果超同阈值时，全文写入 /workspace/saved-<hash>.md|json，结果携带 4k 预览、savedTo 路径和回读提示；相同内容幂等落到同一文件。
• 本地数据库：skills 表（v2）、随会话清理的 files 表（v3）、v4 修复迁移（对历史上 slug 碰撞的技能自动加后缀重命名，最新者保留原名，零数据丢失）。
• 新增 ADR 0015（Agent 沉淀的站点技能）、ADR 0016（fs 工具与文件工作区）、ADR 0017（fetch-first URL 阅读）。

变更

• 固定顶层工具变为 6 个：getDocument、extractImage、navigate、browser、browserRepl、fs（debug 构建仍额外提供 debugger）。
• Agent 指令 v5：目标锚点高效循环（先明确完成标准 → 读最小证据 → 选最短路径 → 一次动作 → 验证），fetch-first URL 策略，文件与技能策略。
• getDocument source:"pdf" 并入 source:"url"（自动识别）；source:"file" 改为按名读取工作区文件，不再接受内联 fileText。
• 产品叙事更新为强大全能的浏览器侧边栏 Agent（CONTEXT/PRODUCT/plan）；停止/恢复与工具轨迹机制不变。
• 商店合规披露补充本地存储的会话文件与站点技能。
• 新增依赖 docx 与 mammoth：纯 JS，动态 import，仅在文档转换时加载。

修复

• 内置种子不再覆盖用户/Agent 沉淀的同名技能；单个种子写入失败不再中断其余种子。
• 技能名 slug 碰撞导致 /skills 路径重复：保存时拒绝冲突，技能文件重命名改为原子操作（先保存后删旧），失败不丢数据。
• 运行时冒烟脚本硬编码的 instructionsVersion 已过时，改为引用共享常量。

Taber v0.2.0

0.2.0 - 2026-07-09

English

Highlights

• Added first public Chrome MV3 release packaging: taber-v0.2.0-chrome-mv3.zip.
• Added subscription login for ChatGPT/Codex and xAI/Grok.
• Reworked the side panel into a product UI with settings, sessions, sources, image previews, toasts, and a compact tool trail.
• Added stronger browser-task flow: target-tab locking, Browser Access setup, streaming AgentHost, and persistent task events.

Added

• ChatGPT/Codex subscription login with PKCE, local token storage, model refresh, and sign out.
• xAI/Grok subscription login with browser OAuth, manual finish-code fallback, token refresh, and sign out.
• OpenAI API key accounts through the Responses API.
• OpenAI-compatible custom providers with presets, model metadata, context windows, and reasoning effort choices.
• Document and visual extraction for pages, selections, PDFs, text files, viewport screenshots, page images, canvas, and background images.

Changed

• Release builds keep debugger out of the extension manifest. The debugger tool is only available in TABER_ENABLE_DEBUGGER=1 local builds.
• Browser tasks require Browser Access setup for website access and User Scripts before page work starts.
• Release install docs now point to the Chrome MV3 zip and unpacked extension flow.

Security

• Provider credentials stay in IndexedDB and are used only for the selected provider.
• Release build verification checks the permission set and provider auth hosts used by ChatGPT/Codex and xAI/Grok login.

中文

重点

• 新增首个公开 Chrome MV3 Release 包：taber-v0.2.0-chrome-mv3.zip。
• 新增 ChatGPT/Codex 与 xAI/Grok 订阅登录。
• 侧边栏重做为产品界面：设置、会话、来源、图片预览、toast 和紧凑工具轨迹。
• 浏览器任务流程增强：target tab 锁定、Browser Access 引导、流式 AgentHost 和持久任务事件。

新增

• ChatGPT/Codex 订阅登录：PKCE、本地 token 存储、模型刷新和退出登录。
• xAI/Grok 订阅登录：浏览器 OAuth、手动 finish code 兜底、token 刷新和退出登录。
• OpenAI API key 账号，走 Responses API。
• OpenAI-compatible 自定义供应商，支持预设、模型元数据、context window 和 reasoning effort。
• 文档与视觉提取：页面、选区、PDF、文本文件、当前视口截图、页面图片、canvas 和 background image。

变更

• Release 构建不在 manifest 中包含 debugger。debugger 工具只在 TABER_ENABLE_DEBUGGER=1 的本地构建里可用。
• 浏览器任务开始前需要完成 Browser Access 的站点访问和 User Scripts 设置。
• Release 安装说明改为指向 Chrome MV3 zip 和加载已解压扩展流程。

安全

• 供应商凭证保存在 IndexedDB，只用于你选择的供应商。
• Release 构建校验会检查权限集合，以及 ChatGPT/Codex 和 xAI/Grok 登录所需 auth host。