做出会替你干活的 AI 智能体。

生成结构化场景计划，在 Lovable 构建工作室，并将批准的合成发送到外部 Remotion 渲染服务。

Remotion 将视频表示为 React 合成。官方 Agent Skills 帮助编码智能体使用它，但并非 Lovable 内置视频生成器。把简报编辑器连接到独立渲染服务，先制作一个 9:16 模板。

Build a Lovable video studio, not a native rendering engine. Save briefs and generate strictly validated scene JSON for a single trusted 9:16 Remotion template. Add scene text and timing editing, asset ownership checks, preview, cost estimate and explicit render approval. Use a server function to call an external authenticated rendering worker; keep keys server-side. Never execute model-generated code. Track queued/rendering/failed/ready states, authenticate callbacks and deduplicate retries by request ID. Store MP4 and thumbnails privately with expiring links. No automatic social publishing. Include tests for long captions, missing assets, duplicate callbacks and cancelled jobs.

不必反复粘贴相同指令。为每项任务选择经过审核的版本化技能，并显示使用版本。

技能包含指令、示例和资源，不等于工具权限或新模型。先加载目录，再按需加载详情。仅建立 Lovable 表并不能保证与外部编码智能体兼容，还需实现对应格式和运行环境。

Build a reviewed skill library for a customer-research agent. Tables: skills, immutable skill_versions, resources, test_cases and runs, scoped to each workspace. Start with one skill that turns interview notes into a structured customer brief with evidence references. Select by catalogue name and description, then load only the chosen reviewed version and relevant resources. Validate selected IDs and user access server-side. Skill instructions cannot override runtime tool permissions. Treat uploaded notes as untrusted data. No arbitrary script execution. Save version and inputs on every run; add evidence review, regression comparison and rollback to a reviewed version.

连接不等于可信。在 Lovable 构建权限目录、变更审核和写入操作审批。

MCP 标准化工具发现和调用，但不保证服务器安全。OWASP 列出工具投毒和定义变更风险。需在服务端按用户权限执行变更与写入审核，面板本身不能强制保护。

Build an MCP tool review workspace for an existing runtime AI agent. Use a server-side MCP client with one documented HTTPS server and its supported OAuth flow. Persist connections per authenticated user; protect tokens server-side. Block private-network targets and redirects. Version tool names, schemas and descriptions; disable new or changed tools pending owner review. Enforce access and allowed tools at runtime, never from model-supplied user IDs. Read tools first. Persist exact write payloads with expiry; require single-use approval bound to the immutable payload and revalidate authorization at execution. Treat descriptions and results as untrusted. Redact logs, close clients and test malicious descriptions, revoked tokens, cross-user requests and duplicate approvals. This is not a security certification.

一个智能体按计划从好几个服务收集价格、挂牌和招聘信息，把一团乱理成一张表，只有真出现变化时才提醒你。

你把一个来源描述一次：它的 URL、页面怎么分页、你需要哪些字段。智能体抓取页面，让模型把字段抽进一个固定 schema，按一个稳定的键去重，存一份快照，并写下对着上一次运行的差异。

Build a scraping and monitoring agent.

Data: table "sources" (name, url, pagination_pattern, selector_hint, fields jsonb, interval_minutes, active). Table "runs" (source_id, status, items_found, error, started_at, finished_at). Table "items" (source_id, raw_key unique per source, title, price numeric, currency, url, published_at, payload jsonb, first_seen_at, last_seen_at). Table "item_changes" (item_id, field, old_value, new_value, created_at).

Server: an edge function that takes a source id, fetches the page with a normal user agent, strips scripts, styles and navigation, trims the text to the listing area, then asks the model to extract rows into a strict JSON schema with temperature zero and null for missing values. Upsert rows into items on raw_key and insert item_changes whenever a tracked field differs. Log every attempt into runs, retry twice with backoff on network errors and respect a per source rate limit. Follow pagination up to a configurable page limit.

Automation: a scheduled job that picks active sources whose interval has elapsed, runs them one by one, then sends a digest email or webhook containing only item_changes created since the previous digest.

UI: sources page with add and edit form, health badge from the last run, run history with error text, items table with search and filters, a change feed and a CSV export button.

Rules: only public pages, honour robots.txt, keep all keys and model calls on the server, enable row level security so each user sees only their own sources and items.

多数团队最先需要的那个智能体：它读你的帮助中心，凌晨三点也能回答客户，还不瞎编政策。

上传文档，把它们切成块，存嵌入，只用检索到的段落加一个来源链接来回答。什么都匹配不上时，智能体转给人工，而不是瞎猜。

Build a documentation support agent.

Data: a table "documents" (title, url, content) and a table "chunks" (document_id, content, embedding vector). A table "escalations" (question, email, created_at, status).

Server: one edge function that embeds the user question, retrieves the six closest chunks, and streams an answer that uses only those chunks. Every answer must end with the source links it used. If similarity is below the threshold, return a short "I do not have this in the docs" answer and insert a row into escalations.

UI: a chat page with streaming responses, message history in the browser, source links rendered as chips under each answer, and an escalation form that appears when the agent cannot answer.

Enable row level security on all tables, keep model calls on the server, and show a clear error state when the request fails.

把那个没人填的联系表单换成一个智能体，问四个问题，自己就把会议约好。

智能体做一场简短的结构化访谈，对着你的标准给线索打分，把结果写进数据库，并给销售发一份小结。

Build a lead qualification agent widget.

Behaviour: the agent asks at most five questions, one at a time, to learn company size, budget range, timeline and the problem being solved. It never asks two questions in one message and never asks for data it already has.

After the interview, produce a structured result with fields: score, tier, summary, next_step. Keep the schema flat and unconstrained, state the ranges in the prompt text, and clamp values in code.

Storage: insert into a "leads" table with contact details, transcript, score, tier and created_at. Send a notification email with the summary.

UI: a floating chat bubble on the landing page, mobile friendly, with a typing indicator and a final card showing the booked next step.

一个多步智能体，会搜索、阅读、比较，交回一份带出处的简报，而不是一段自信的话。

定义三个工具：搜索、抓取页面、保存发现。让智能体循环，直到它有了足够的证据，然后在 UI 里渲染工具时间线，好让用户能审计每一步。

Build a research agent with tool calling.

Tools: web_search(query), fetch_page(url), save_finding(claim, source_url, confidence). Give each tool a narrow input schema and a short description.

Loop: allow at least fifty steps, stop when the agent has five saved findings or decides the question is answered. Never let the agent answer without at least two independent sources.

Output: a brief with a one paragraph answer, a bullet list of findings, and a sources section. Save the brief to a "reports" table.

UI: stream the answer, render every tool call as a timeline row with its input and a compact result, and add a copy button plus markdown export.

粘一篇文章，拿回邮件通讯、推文串、图文脚本和视频钩子，全用你自己的口吻。

存一个带规则和禁用短语的语气档案，然后在单独的调用里生成每一种格式，好让输出保持锋利、而不是糊成一团。

Build a content repurposing agent.

Input: a long article plus a saved voice profile (tone rules, favourite words, banned phrases, audience).

Generate four outputs in separate calls so each one is focused: a newsletter intro, a social thread of six posts, a carousel script of seven slides, and three video hooks.

Rules for every output: no em dashes, no emoji, no filler phrases like "in today's fast paced world", short sentences, concrete nouns.

UI: one input panel on the left, four result cards on the right, each editable inline with a copy button and a regenerate button that only regenerates that card.

每个创始人都被同样的四十封邮件淹没。这个智能体把它们分类、起草回复，发出去之前先问你。

把每条消息分类到一小组意图里，附上一个建议动作，任何离开这个应用的东西都要求明确批准。

Build an inbox triage agent.

For each incoming message, classify it as one of: sales, support, partnership, invoice, noise. Add urgency low, medium or high, and a one sentence reason.

Draft a reply for every message that is not noise, using saved templates as the base and matching the sender's language.

Any action that sends an email or changes external state must require approval in the UI first. Show the draft, the intent and the reason on an approval card with approve, edit and discard.

UI: a triage queue grouped by intent, keyboard shortcuts for approve and skip, and a daily digest view of what was handled.

诊所、美发店和工作坊到现在还漏掉一半来电。一个语音智能体第一声就接起来，直接把预约写进日历。

进来时语音转文字，中间一段简短的脚本化对话，出去时文字转语音，再加一条来电者能核对的确认消息。

Build a voice booking agent.

Flow: transcribe the caller audio, run a slot filling dialogue that collects service, preferred day, preferred time and phone number, confirm the details out loud, then create the booking.

Never invent availability. Check the bookings table first and offer the two closest free slots when the requested time is taken.

Confirm by reading the booking back and sending a text summary. Store the full transcript with the booking for review.

UI: an operator dashboard with today's calls, transcripts, created bookings and a flagged list of calls the agent could not finish.

多数流失发生在头十分钟。一个知道用户卡在哪的智能体，比任何气泡引导都更能救激活。

把用户已完成的步骤和当前所在的界面喂给智能体，让它建议唯一的下一个动作，并给它一个真能执行那个动作的工具。

Build an in-app onboarding copilot.

Context: pass the current route, the list of completed setup steps and the last three user events into the system prompt on every call.

Behaviour: always propose exactly one next action, explain in one sentence why it matters, and offer to do it. Never list five options.

Tools: create_project, invite_teammate, connect_data_source. Mark every tool as needing approval, and show the approval card inline in the chat.

UI: a slide-over panel available from every page, a progress checklist at the top, and an admin view showing where users get stuck most often.

用大白话问一个问题，拿到一张图和它背后的查询。愿意把查询亮出来，正是让人信它的地方。

把 schema 给智能体，把它限制在只读视图上，让它写查询、在服务器端跑，再把结果渲染成一张图表，SQL 就显示在下面。

Build a data analyst agent over a read-only schema.

Safety: expose only read-only views, reject anything that is not a select, cap rows returned, and run every query server side with a timeout.

Behaviour: the agent writes the query, explains in one sentence what it measures, then picks the chart type that fits the shape of the result.

UI: a question box with example questions, a chart, a collapsible SQL block, and a saved questions list that can be rerun with one click.

Handle empty results and query errors explicitly with a plain language message and a suggested fix.

不是又一个洗稿工具。这个智能体读当前的搜索结果页，找出所有人都漏掉的，给一个配得上点击的页面写简报。

收集排名靠前的结果，抽出它们的标题，把共享的子话题聚簇，输出一份大纲，带上没人覆盖的角度加内部链接目标。

Build an SEO brief agent.

Input: a target keyword and the site the page will live on.

Steps: search the keyword, fetch the top eight results, extract their headings and word counts, cluster the shared subtopics, and list the questions none of them answer well.

Output a brief with: search intent in one line, target title and description, an H2 outline, entities to mention, three internal link targets from the given site, and the differentiating angle.

UI: a single input, a progress list of the steps as they run, and the finished brief in a copyable panel with markdown export.

筛选器让人走开。一段关于他们到底需要什么的短对话，能把他们带到商品页、准备下单。

智能体问用途和约束，查实时目录，返回三个选项，配一份对权衡的朴素比较。

Build a shopping assistant agent for an online store.

Tools: search_products(query, filters), get_product(id), check_stock(id).

Behaviour: ask at most three questions about use case, budget and any dealbreaker, then recommend exactly three products. For each one, give the reason it fits and the honest downside. Never recommend an out of stock item.

UI: chat with product cards inline, each card showing image, price, stock status and an add to cart button. Keep the conversation on the product page as a side panel.

Log every conversation with the products shown and whether the visitor added to cart, so the store can measure assisted revenue.

两百份申请，每个候选人一张诚实的评分表，还有每个决定到底为什么这么定的记录。

解析简历，对着你写的标准给它打分，通过邮件问两个澄清问题，把短名单连同推理一起交给一个人。

Build a recruiting screening agent.

Input: a job description with must have and nice to have criteria, plus uploaded resumes.

For each candidate produce: a score per criterion, an overall recommendation, the strongest evidence quoted from the resume, and the biggest open question.

Never infer age, gender, nationality or anything the criteria do not mention, and never reject a candidate automatically. The agent proposes, a human decides.

UI: a candidate table sorted by score with expandable scorecards, a compare view for the top five, and a one click email that asks the open question.

上传一张账单，拿到分类、趋势，和一句关于你忘了的那个订阅的、让你不太舒服的话。

解析文件，给交易归类，检测周期性扣费，让用户就自己的数字问跟进问题。

Build a personal spending agent.

Input: a CSV bank statement uploaded by the user.

Processing: normalise dates and amounts, categorise every transaction using rules first and the model only for the leftovers, and detect recurring charges by matching merchant and cadence.

Output: a summary with total by category, month over month change, the three largest changes explained in one sentence each, and a list of subscriptions with their yearly cost.

UI: upload area, a category chart, a transactions table with inline category editing that retrains the rules, and a chat box for follow up questions about the uploaded data only.

Keep all data private to the signed in user with row level security.

一个聊天窗口，读你的 CRM、你的文档和你的账单，然后把那件谁都不想开五个标签页去做的跨工具琐事做掉。

Model Context Protocol 把工具访问变成了一个标准。不用为每个服务写一个单独的集成，智能体拿到一份类型化工具的清单，在运行时挑对的那一个。你维护一个连接的注册表、一份按用户角色的权限映射，以及一份智能体做过的每一次调用的日志。

Build an internal tool-calling agent with a typed tool registry.

Data: connections (user_id, service, label, status), tools (name, description, args_schema, mode read or write), tool_calls (tool, args, result, duration_ms, cost), approvals (tool_call_id, status).

Server: one function that returns the tools allowed for the current user role, one function that executes a tool after validating arguments against args_schema, and one that records the call.

Agent loop: the model receives the allowed tool list, chooses at most one tool per turn, and must explain in one sentence why it chose it.

Safety: any tool marked write requires an approval card showing the exact arguments before it runs. Never execute a write tool without a recorded approval row.

UI: chat panel, a connections page with a test-run button, and a run timeline showing each tool call with duration and cost.

Protect every table with row level security so a user only ever sees their own connections and calls.

今年被搜得最多的智能体。它拿一份公司名单，找到该联系的理由，写一条不像模板的消息，再把回复放上你的日历。

价值不在于发更多邮件，在于那个调研的步骤。对每一家公司，智能体收集几个公开信号，对着你的理想客户画像给契合度打分，把标准线以下的都丢掉，然后才起草。每一份草稿都要等一个人点一下才发出去。

Build an AI SDR agent focused on research quality, not send volume.

Data: accounts (name, domain, size, notes), signals (account_id, type, text, source_url, found_at), fit_scores (account_id, score, reason), drafts (account_id, subject, body, status), replies (draft_id, category, next_step).

Pipeline: import accounts from CSV, research each account and store at least one signal with a source URL, score fit from 0 to 100 against a stored ideal customer profile, and draft only for accounts above the threshold.

Writing rule: every draft must quote exactly one stored signal and stay under 90 words. If no signal exists, skip the account and say why.

Human control: drafts land in an approval queue with inline editing. Nothing sends without an approved row and a per-day send cap.

Replies: classify each reply as interested, not now, or not a fit, and surface interested ones with a suggested meeting time.

UI: pipeline board, an account page showing signals with source links, and a draft queue with approve, edit and skip.

十二个脆弱的场景、三个工具，没人记得哪个才是发发票的那个。把它们收进一个应用，每次运行都看得见，每次失败都有名字。

确定性的步骤保持确定性。模型只用在一个决定需要判断的地方：给这个分类、选收件人、决定它能不能等。别的一切都是排程上的普通代码，那正是让整件事既便宜、又以最好的方式无聊的原因。

Build a workflow runner app where an AI step is used only for judgement.

Data: workflows (name, schedule, active), steps (workflow_id, position, kind rule or model, config), runs (workflow_id, started_at, status), run_steps (run_id, step_id, input, output, duration_ms, error), dead_letters (run_id, payload, reason).

Execution: run steps in order, store input and output of every step, retry a failed step three times with growing delay, then write it to dead_letters and stop the run.

Model steps: must return exactly one value from a fixed list defined in the step config, with a one sentence reason stored alongside.

UI: workflow list with last run status, a run timeline showing each step with duration and output, a manual run button, and a dead letter screen with a retry action.

Keep secrets in server functions only and protect every table with row level security.

一家诊所、一家美发店或一家修理铺，会漏掉三分之一的来电。每一个都是一个付费客户，只是刚拨了名单上的下一个号码。

智能体接起来，回答每个人都问的那五个问题，查日历，预约或记下留言，并给老板发一份小结。诀窍在那个兜底：当来电者听着不耐烦、或者问了脚本之外的东西，它就提出转人工、然后停止说话。

Build an AI receptionist for a local business.

Data: business_profile (hours, services, prices, address, five FAQ pairs), calls (started_at, caller, outcome), transcripts (call_id, role, text, confidence), bookings (service, slot, caller, status hold or confirmed), messages (call_id, text, callback_number).

Behaviour: greet, identify the request, answer only from business_profile, check available slots, place a hold, confirm, then release or confirm the hold.

Limits: answers stay under two sentences. If the model cannot answer from business_profile, or the caller sounds frustrated, offer a callback and store the number instead of guessing.

After the call: generate a summary with outcome, next step and a link to the transcript, and send it to the owner.

UI: a text chat to test the same logic without voice, a call log with transcripts, a bookings calendar, and an editable business profile page.

最快搭出来、也最好演示的智能体。粘一个念头，拿到七个钩子、七条脚本和文案，全用你自己的口吻。

语气就是产品。智能体存下十份你实际怎么写的样本，把模式抽取一次，应用到每一个脚本上。没有那一步，你得到的就是别人都在发的同一个通用钩子。

Build a short video script agent that writes in the user's own voice.

Setup: the user pastes ten samples of their own writing. Extract a voice profile once with tone, sentence length, favourite structures and banned words, and store it.

Input: one idea in a sentence.

Output: seven hooks scored for curiosity and clarity, the best three expanded into 40 second scripts using hook, tension, payoff and call to action, plus a caption and five tags for each.

Rules: no generic openers, no words from the banned list, and every script must be readable out loud in under 45 seconds.

UI: idea input, a scored hook list, a script editor, and a weekly calendar where scripts can be dragged between days.

进去一份说明，出来一条成片：脚本、分镜、配音、字幕，还有给不同投放位切的三个版本。大家低估的是分镜清单，不是渲染。

智能体分阶段工作，而不是一个巨大的提示词。它先写出报价和承诺，然后一份逐镜头的分镜表、每个镜头带时长，然后配着那些时长掐点的配音文本，然后它把片段拼起来、把字幕烧进去。每个阶段都被保存，所以你能重写一个场景、而不用重新生成整支广告。

Build an ad video agent that turns a product brief into a finished commercial.

Data: briefs (product, buyer, pain, promise, proof), scripts (brief_id, angle, status), shots (script_id, index, seconds, visual, voice_over, on_screen_text, clip_url), renders (script_id, aspect, url, status).

Flow, one stage at a time and each stage saved before the next:
1. From the brief, propose three angles with a one line promise each. Wait for the user to pick one.
2. Build a shot list as rows with index, seconds, visual, on screen text. Total duration must equal the chosen placement length of 15, 30 or 60 seconds.
3. Write the voice over per shot, sized to that shot's seconds at about 2.5 words per second.
4. Attach a clip to each shot, then queue renders for 9:16, 1:1 and 16:9 with burned in word level subtitles.

Rules: one promise per ad, the hook lands in the first two seconds, no claim that is not backed by the proof field, and never exceed the placement length.

UI: brief form, angle picker, an editable shot table where changing seconds re-times only that shot's voice over, a render queue with status, and a version list to compare cuts side by side.

付费社媒不是败在剪辑，是败在头两秒。这个智能体每个产品产出二十个钩子，留下扛过打分的那些，再变成能直接开拍的达人简报。

输出不是一个脚本，是一套拍摄套件：钩子句、创作者在每一拍里拿什么做什么、要念出来的那句话、说明文字和缩略图画面。每个变体都带一个追踪名，好让广告结果能被粘回来，智能体就学会哪个钩子家族赢。

Build a UGC ad script agent that produces shot ready creator briefs and learns from results.

Data: products (name, promise, proof, audience), hooks (product_id, family, line, curiosity_score, clarity_score, claim_risk), variants (hook_id, beats json, caption, tags, tracking_name), results (tracking_name, spend, views, clicks, conversions).

Hook families: problem, contrast, mistake, result, demo, story. Generate three hooks per family, score each from one to five on curiosity and clarity, flag any claim not supported by the proof field, keep the top eight and discard the rest.

For each surviving hook produce four beats of five seconds: hook, problem, demo, call to action. Every beat states what is in frame, what the creator does with their hands, and the exact line spoken. Add a caption, three tags and a tracking name of the form product-family-number.

Results loop: accept a CSV of results by tracking name, compute cost per click per hook family, and weight the next generation toward the winning families.

UI: product form, a scored hook table, a beat editor, a print view of the creator brief, and a results dashboard by hook family.

录一段乱糟糟的五分钟走查，拿回一条 60 秒的演示，配好文字旁白、每个要紧点击都放大、字幕也干净。毁掉演示的是冷场，而这正是智能体去掉的东西。

转写只是入口。智能体把转写映射到录像里的那些时刻，去掉静音和重复的尝试，把啰嗦重写成简短的、以好处为主的旁白，然后在发生点击的时间戳上放置放大和高亮标记。用户编辑的是一条片段的时间线，不是一个视频文件。

Build a product demo video agent that turns a raw screen recording into a short polished demo.

Data: recordings (file_url, duration, target_length), words (recording_id, text, start, end), segments (recording_id, start, end, task, narration, importance), markers (segment_id, timestamp, type zoom or highlight, region), exports (recording_id, aspect, url, status).

Pipeline: transcribe with word level timestamps, remove silences longer than 0.6 seconds and repeated sentence starts, group the rest into segments where one segment equals one task the product performs, rewrite each segment into a single benefit sentence, score segments by importance and keep only enough to fit the target length.

Then place a zoom marker at every click timestamp and a highlight on the region that changed, render with word level subtitles and export 16:9 and 9:16 from the same edit.

Rules: narration describes the outcome, not the interface. Never say click here. Total length must not exceed the target. Keep every cut reversible by storing segments rather than rewriting the source file.

UI: upload page, a segment timeline where segments can be reordered, muted or restored, a narration editor with live duration, a marker overlay on the preview, and an export panel per aspect ratio.

把一个复杂目标交给协调智能体。它拆分任务、分派给专门智能体，并在独立审核通过后才交付。

这不是机器人群聊。每项任务都有负责人、输入约束、预算、截止时间和验收测试。失败一次后转交人工，避免无限循环。

Build a multi-agent project workspace with a coordinator, researcher, executor and independent reviewer.

Data: projects, tasks, runs, artifacts and approvals. A task must include objective, input artifact IDs, allowed tools, budget, deadline, dependencies and a measurable acceptance test.

Flow: the coordinator creates a dependency graph. Specialists can run only when dependencies are complete. Every output becomes an immutable artifact with sources and cost. The reviewer receives the acceptance test and artifacts, returns pass or fail with exact reasons, and cannot edit the result. A failed task gets one targeted retry, then requires human intervention.

Safety: cap parallel jobs and total spend, stop on missing evidence, never expose chain of thought, and require a human approval card before final delivery.

UI: project brief, dependency graph, live task timeline, artifact viewer, cost meter, review report and final approval queue.

智能体可以搜索、比较并填写重复表单，但购买、发消息、提交和删除始终需要用户明确确认。

智能体以结构化浏览器操作进行规划，每一步后检查页面并记录截图和网址。敏感字段和不可逆按钮会生成操作预览并等待确认。

Build a safe browser computer-use agent for one narrow web workflow.

Data: jobs, browser_steps, observations and approvals. Every step is a typed action: navigate, inspect, click, type or extract. Store the expected page change, actual URL, title, screenshot and observation.

Rules: use only allowlisted domains, stop on login challenges or unexpected states, enforce a hard step limit, and never put secrets or payment credentials in model context. Payment, send, submit, publish, delete and sensitive-data entry always require approval.

Approval card: show the current screenshot, destination, exact payload and consequence. Execute only after a fresh user confirmation. A visible stop control must close the session and invalidate every queued action.

UI: job brief, live browser view, action timeline, approval queue, masked audit log and stop control.

在生成和发布之间加入验证层，检查来源、规则、隐私和格式，然后通过、阻止或附带具体原因转交人工。

先运行便宜且可重复的确定性规则，独立模型只根据版本化政策评估模糊检查。高风险或无证据内容不能自行发布，每个决定都保存规则、证据和政策版本。

Build an AI quality and safety agent that gates every generated result before publication.

Data: versioned policies, test_cases, evaluations and incidents. Run deterministic checks first for schema, required fields, length, PII patterns, links and prohibited claims. Send only ambiguous checks to a separate evaluator with the exact policy version and source evidence.

The evaluator returns structured JSON: decision pass, review or block; risk score; rule IDs; evidence spans; and recommended correction. The producing agent must not have access to publish unless the decision is pass.

Human review: show original input, generated output, failed rule, evidence and correction side by side. Every confirmed incident becomes a permanent regression test.

// APPEND-ANCHOR
UI: policy editor with versions, live evaluation queue, pass-review-block dashboard, incident log and metrics by model and policy version.

输入一份简报，输出一条成品 Reel：开场钩子、脚本、配音、字幕、B-roll 计划和封面，组成可重复的流水线，而不是靠运气。

智能体接收主题和受众，写出三个开场钩子变体，把最佳方案扩展成 30 秒脚本，生成配音和带时间码的字幕，为每句台词规划 B-roll，并导出剪辑师或视频模型可执行的镜头清单。每一步都保存结果，渲染失败不会拖垮整个流程。

Build a reels production agent that turns one brief into a finished vertical video plan.

Data: briefs, hooks, scripts, voiceovers, captions, shot_lists and renders. From a brief (topic, audience, tone, call to action, banned claims) generate three hooks under 8 words, score them, and expand the winner into a 30-second script: hook, three beats, payoff, call to action.

Generate voiceover with word-level timestamps, captions synced to the audio, and a b-roll instruction per script line with duration, framing and source. Assemble a shot list and a cover frame, then push the render to a queue with retries.

Human control: every step's artifact is editable before the next step runs. Publishing requires explicit approval.

UI: brief form, hook comparison, script editor with timing, shot list board, render queue with status, and a cost-per-reel dashboard.

智能体每天早上监测你所在领域的趋势音效、格式和钩子，在竞争对手喝完咖啡之前交出一份带现成脚本的 Reels 创意排名清单。

定时采集器拉取你领域的新帖子，提取音效、话题标签、钩子模式和互动增速，评分器按新颖度和账号匹配度排名机会。输出每日摘要：五个 Reels 创意，每个都有钩子、30 秒脚本骨架和支撑它的趋势证据。

Build a reels trend research agent that answers one question every morning: what should I post this week.

Data: trends, sounds, hashtags, posts, ideas and digests. The user defines a niche with 10 to 20 seed accounts, competitor hashtags and banned topics. A scheduled collector pulls fresh posts each morning and stores sounds, hashtags, captions and view counts.

Extract the first line of each video as its hook and cluster repeated hook patterns. Compute engagement velocity as views per hour since posting. Score every trend by novelty in the niche, velocity and fit to the user's content pillars.

Generate a daily digest of five reel ideas. Each idea has a hook, a three-beat 30-second script skeleton and links to the source posts as evidence. Deliver the digest to email or Telegram at a fixed hour.

UI: niche setup, trend board with velocity charts, idea cards with script skeletons, digest history, and a feedback loop that records which ideas were posted and how they performed.