跳转到正文
skillsafebeta
← 报告库
用途分类 / 其他用途

Higgsfield Generate Skill 安全审计

作者说它能做什么（原文）

|

第三方安全检查结论

先别安装或运行

已检查文件
13
发现的风险
7
会不会运行危险命令？检查是否下载程序后直接运行、让他人远程控制电脑，或藏起要运行的命令。发现 2 项风险
高风险

下载后直接运行外部程序

原文依据：2 处
发现了什么

这一行把下载到的内容直接交给电脑运行，中间没有让你检查内容的步骤。

为什么需要注意

一旦运行，下载来的程序就可能读取或修改你的文件。即使网址没变，对方也可能换掉里面的程序。

当 `higgsfield` 不在 PATH 时，这是一条生效的安装指令：它把 GitHub 当前返回的脚本直接交给 shell 执行。脚本未固定到提交或校验和；若远端内容或传输链路被攻破，代码将以当前用户权限运行，可能访问或更改其可访问的数据。用户可要求作者提供固定版本、校验和及先下载审阅的安装方式。

SKILL.md:37来自说明文档打开原文件
   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh   ```
查看另外 1 个位置
SKILL.md:35来自说明文档打开原文件
1. If `higgsfield` is not on `$PATH`, install it:   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh   ```2. If `higgsfield account status` fails with `Session expired` / `Not authenticated`, ask the user to run `higgsfield auth login` (interactive) and wait for confirmation.
高风险

引导直接执行联网下载的安装脚本

原文依据：2 处
发现了什么

当 CLI 不存在时，Skill 要求把 GitHub 上当前返回的脚本直接传给 shell，未固定版本、提交哈希或校验和，也没有先供用户审阅。

为什么需要注意

如果下载地址、发布账户、网络链路或脚本内容被篡改，脚本会以代理当前权限执行任意命令，可读取或修改用户文件与凭据。

候选结论与实际引导一致：在缺少 CLI 时，Skill 要求执行来自 `main` 分支的联网脚本，没有可见的版本固定、哈希校验或审阅步骤。这不是示例或警告，而是“Before any other command”的启动流程。用户可要求可验证的固定发行包，并限制安装过程的文件和网络权限。

SKILL.md:33来自说明文档打开原文件
Before any other command:1. If `higgsfield` is not on `$PATH`, install it:   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh   ```2. If `higgsfield account status` fails with `Session expired` / `Not authenticated`, ask the user to run `higgsfield auth login` (interactive) and wait for confirmation.
查看另外 1 个位置
SKILL.md:31来自说明文档打开原文件
## Step 0 — BootstrapBefore any other command:1. If `higgsfield` is not on `$PATH`, install it:   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh   ```2. If `higgsfield account status` fails with `Session expired` / `Not authenticated`, ask the user to run `higgsfield auth login` (interactive) and wait for confirmation.
会不会泄露文件和密钥？检查是否发送含密码或密钥的文件，以及代码里是否直接写了密钥。发现 2 项风险
高风险

本地照片、视频和声音会自动上传到第三方服务

原文依据：5 处
发现了什么

Skill 明确说明媒体路径会被自动上传，并给出了上传创始人照片、声音参考和广告视频的流程。它没有要求在上传前说明接收方、保留期限或取得涉及人物的同意。

为什么需要注意

人脸、声音、未发布广告、客户素材或含元数据的文件会离开本机，可能受第三方存储、处理和账户保留规则约束。

本地路径会由 CLI 自动上传；流程明确涵盖照片、视频和声音，包括创始人头像和用于分析的广告视频。上传是所请求生成功能的正常组成部分，并非暗中窃取，但会把可能含人脸、声音或商业素材的数据交给 Higgsfield。可见说明未在上传前要求确认接收方、保留期或人物授权。用户可要求每次上传前确认，并只提供已获授权、去敏后的素材。

references/media-inputs.md:5来自说明文档打开原文件
## Path or UUID — both workEach media flag accepts either a local file path or a UUID. The CLI auto-uploads paths before submission and auto-detects whether a UUID is an upload id (from `higgsfield upload create`) or a previous job id.
查看另外 4 个位置
references/marketing-avatars.md:23来自说明文档打开原文件
```bashID=$(higgsfield upload create founder.png)URL=$(higgsfield upload create founder.png --json | jq -r .url)   # if you need cloudfront URLhiggsfield marketing-studio avatars create --name "Founder" --image $ID --image-url $URL```
references/media-inputs.md:19来自说明文档打开原文件
# Video analysis — CLI uploads the file, Virality Predictor returns a text score/report plus an Open report link.# The output is text, but the task is still video analysis.higgsfield generate create brain_activity --video ./ad.mp4 --wait```
references/media-inputs.md:103来自说明文档打开原文件
higgsfield generate create seed_audio \  --prompt "same voice, calmer delivery" \  --audio-references ./voice.wav \  --wait```
references/marketing-avatars.md:21来自说明文档打开原文件
## Creating a custom avatar```bashID=$(higgsfield upload create founder.png)URL=$(higgsfield upload create founder.png --json | jq -r .url)   # if you need cloudfront URLhiggsfield marketing-studio avatars create --name "Founder" --image $ID --image-url $URL````--image-url` is the cloudfront URL from the upload. Required by the API.
中风险

上传素材会形成可复用的账户对象，但未说明清理方法

原文依据：6 处
发现了什么

广告参考会在后端处理并存储，头像、产品和品牌资料也会作为可复用实体创建。所给说明只有创建、列出和读取流程，没有可见的删除或保留控制。

为什么需要注意

即使单次生成已经结束，人脸照片、产品素材、参考视频及抓取的品牌资料仍可能留在账户中，扩大后续误用或账户泄露时的暴露范围。

广告参考被明确描述为由后端处理、存储并可日后调用；头像和产品也会创建为账户中的实体。可见文档提供创建、列出和读取流程，却未提供删除命令或保留期限，因此用户可能留下长期可复用的个人或商业素材。这不证明服务实际上没有删除能力。用户可在上传前要求作者说明保留、访问范围和删除办法，并避免创建不必要的持久对象。

references/marketing-ad-references.md:3来自说明文档打开原文件
Ad references are reusable inspiration videos a user wants to model new ads after — typically tied to a specific avatar and/or product. The backend processes the input video and stores a reusable reference the user can recall later.
查看另外 5 个位置
references/marketing-ad-references.md:52来自说明文档打开原文件
The backend kicks off processing asynchronously. Newly created references start in `status: queued` then move to `in_progress` and finally `completed` (or `failed`).
references/marketing-brand-kits.md:3来自说明文档打开原文件
A brand kit captures a brand's identity (name, logo, hero images, colours, fonts, tone, products) and is reused across image generations to keep visuals on-brand.This is a Marketing Studio metadata import for ad generation, not a visual-identity design workflow. For a new or extended logo, palette, typography system, applications, or editable Brandbook, route to `higgsfield-brandkit` instead. Use this page only to fetch website metadata into a Marketing Studio brand-kit ID.A kit is created by handing in a website URL — it is being fetched and fields are being filled.
references/marketing-products.md:3来自说明文档打开原文件
Two ways to register a product: URL fetch (auto-imports title, description, images) or manual (provide your own).
references/marketing-ad-references.md:54来自说明文档打开原文件
## Discover```bashhiggsfield marketing-studio ad-references listhiggsfield marketing-studio ad-references list --jsonhiggsfield marketing-studio ad-references list --size 50 --cursor <cursor>```
references/marketing-ad-references.md:70来自说明文档打开原文件
## Inspect```bashhiggsfield marketing-studio ad-references get <id>higgsfield marketing-studio ad-references get <id> --json```
会不会删除文件或一直在后台运行？检查是否大范围删除文件、改写磁盘，或设置自动启动。未发现风险
会不会绕过安全保护？检查是否跳过网站安全验证、开放过多文件权限，或取消操作前的确认。未发现风险
会不会误导 AI 或隐藏内容？检查工作说明是否要求 AI 忽略你的指令、干扰检查结果，或夹带看不见的文字。未发现风险
会不会偷偷改推广链接或收款方？检查是否强制替换推广链接或收款对象，同时要求隐瞒更改。发现 3 项风险
中风险

默认直接提交可能消耗额度的高质量任务，不主动展示成本

原文依据：4 处
发现了什么

Skill 明确要求除非用户询问，否则不预估成本，并优先选择质量默认模型；随后的一般流程会直接创建并等待任务。

为什么需要注意

用户可能在不知道预计额度、模型价格或批次成本的情况下产生收费任务，尤其是 4K、长视频或多张广告输出。

Skill 明确要求默认不预估费用、优先质量模型，并在通用流程中直接创建和等待任务。另有成本查询命令，说明提交前估算是可行的，但仅在用户主动询问时使用。因此，未明确预算的请求可能消耗额度，且高质量默认项可能更贵。用户可限制为“提交前总是显示积分成本并确认”，并设置批量大小、分辨率和重试上限。

SKILL.md:48来自说明文档打开原文件
4. Don't batch-ask. Pick a sane default model and ask one thing at a time only if genuinely missing.5. Don't pre-estimate cost or optimize for cheaper models unless the user asks. Prefer the quality default first.6. Pass `--wait` to `generate create` so the command blocks until done and prints the result URL itself. Avoid the two-step `create` → `wait` pattern.
查看另外 3 个位置
SKILL.md:124来自说明文档打开原文件
3. **Validate quickly.** If unsure of params, run `higgsfield model get <jst> --json` once and pass only what's needed. Validate the preferred model before falling back to an older one. Use schema defaults otherwise. The server returns `adjustments` for non-fatal coercions (e.g. `aspect_ratio=99:99` → closest match) and a structured error for invalid declared-param values.4. **Submit and wait in one shot.** `higgsfield generate create <jst> [--prompt "..."] [media flags] [param flags] --wait`. Blocks until terminal status and prints the result on stdout. Tunables: `--wait-timeout 20m` (default 10m), `--wait-interval 5s` (default 3s). Virality Predictor does not need a prompt; pass `--video`.5. **Deliver.** For generated media and 3D assets, send the primary result URL plus a one-line summary (model, duration if video; GLB/asset URL for 3D). For Virality Predictor, deliver the scores, business interpretation, and the Open report link. Do not surface Virality Predictor `.glb`, `.bin`, or region-table internals in normal chat output.
references/marketing-dtc-ads.md:67来自说明文档打开原文件
| `--quality` | no | `low` | `low, medium, high` || `--batch-size` | no | `1` | 1..20 || `--media` | no | — | Repeatable, ≤14. Format: `<media_input_id>[:role]` (default role `image`). || `--avatar` | no | — | Max 1. Format: `<avatar_id>[:type]` (default type `preset`; use `custom` for user-uploaded avatars). || `--product` | no | — | Max 1. Product UUID from `marketing-studio products list`. || `--folder-id` | no | — | Folder placement. || `--from-file` | no | — | JSON file with the params shape. Flags merge over file values (flags win). || `--cost-only` | no | `false` | Print credit cost; do not create a job. || `--wait` | no | `false` | Poll until terminal status, then print the result URL. || `--timeout` | no | `5m` | Wait timeout. |
references/troubleshooting.md:30来自说明文档打开原文件
## Cost`higgsfield generate cost <jst> ...` returns credit estimate without submitting. Useful when the user asks "how much will this cost?".
中风险

无效参数可被自动改成近似值后继续生成

原文依据：4 处
发现了什么

文档称不支持的封闭枚举值会被钳制为最接近的允许值并继续，通用流程只要求接收 `adjustments`，没有要求在付费提交前让用户批准变化。

为什么需要注意

生成结果的比例、时长或其他参数可能不同于用户要求，同时任务仍消耗额度；对有严格交付规格的广告或客户素材尤其有影响。

文档明确说封闭枚举中的不支持值可能被改为最接近值并继续，而通用流程随后直接提交任务。这样可能产生并计费一个与用户指定比例或其他参数不同的结果；可见流程只报告 `adjustments`，没有要求提交前批准。并非所有错误都会这样处理，未知参数等仍会终止。用户可要求出现任何调整时停止，并在重新提交前展示最终参数。

SKILL.md:123来自说明文档打开原文件
2. **Pass media inputs straight to flags.** Media flags accept a local file path **or** a UUID. CLI auto-uploads paths and auto-detects job vs upload for UUIDs. No need to pre-upload. Each model declares accepted media roles or `*_references` params — see `references/media-inputs.md`.3. **Validate quickly.** If unsure of params, run `higgsfield model get <jst> --json` once and pass only what's needed. Validate the preferred model before falling back to an older one. Use schema defaults otherwise. The server returns `adjustments` for non-fatal coercions (e.g. `aspect_ratio=99:99` → closest match) and a structured error for invalid declared-param values.4. **Submit and wait in one shot.** `higgsfield generate create <jst> [--prompt "..."] [media flags] [param flags] --wait`. Blocks until terminal status and prints the result on stdout. Tunables: `--wait-timeout 20m` (default 10m), `--wait-interval 5s` (default 3s). Virality Predictor does not need a prompt; pass `--video`.5. **Deliver.** For generated media and 3D assets, send the primary result URL plus a one-line summary (model, duration if video; GLB/asset URL for 3D). For Virality Predictor, deliver the scores, business interpretation, and the Open report link. Do not surface Virality Predictor `.glb`, `.bin`, or region-table internals in normal chat output.
查看另外 3 个位置
references/model-catalog.md:196来自说明文档打开原文件
These are model-specific. The CLI clamps unsupported values to the nearest allowed one (with a `Note: adjustments applied` warning) when the model declares a closed set. When in doubt:
references/model-catalog.md:217来自说明文档打开原文件
- **Adjustments** — a non-fatal coercion. E.g. you passed `aspect_ratio=99:99` and the model accepts a closed set; the CLI picks the closest match and continues. The adjustments map is included in the response.- **Validation error** — a fatal mismatch. E.g. an unknown declared parameter, or a media role the model doesn't accept. The CLI returns an error and does not submit.
references/model-catalog.md:215来自说明文档打开原文件
The CLI reports two kinds of feedback:- **Adjustments** — a non-fatal coercion. E.g. you passed `aspect_ratio=99:99` and the model accepts a closed set; the CLI picks the closest match and continues. The adjustments map is included in the response.- **Validation error** — a fatal mismatch. E.g. an unknown declared parameter, or a media role the model doesn't accept. The CLI returns an error and does not submit.
中风险

把模型生成的传播力评分描述为“客观”商业依据

原文依据：3 处
发现了什么

Skill 指示把 Virality Predictor 解释为创意测试的“客观注意力代理”，并据脑区分数推断刺激、专注和走神，但可见材料没有提供验证范围、误差或适用人群。

为什么需要注意

用户可能据此淘汰广告、调整预算或评价创意人员，把未经说明的模型推断当成客观测量，造成错误商业决策。

生效说明要求把该评分解释成创意测试的“客观注意力代理”，并把较高脑区分数和较低 Default Mode 分别关联到专注与较少走神。可见材料没有给出验证研究、误差范围或适用人群，用户可能对广告或内容作出过度确定的商业判断。这只支持“可能误导决策”的风险，不证明评分错误。用户可要求披露方法和验证范围，并把结果限制为辅助信号而非唯一决策依据。

SKILL.md:278来自说明文档打开原文件
Use Virality Predictor (`brain_activity`) when the user wants to evaluate a finished video as a business creative: hook strength, virality potential, attention, retention, or how well the content/product holds focus and minimizes distraction. Treat "Virality Predictor" as the customer-facing feature name; `brain_activity` is only the CLI/job_set_type.
查看另外 2 个位置
SKILL.md:284来自说明文档打开原文件
The result is text, not a generated image/video. Report the overall score, peak hook second, sustain score, strongest/weakest regions, and report URL if present. Interpret it as an objective attention proxy for creative testing: higher Visual/Auditory/Language/Attention scores suggest stronger stimulus and focus; lower Default Mode is better because it suggests less mind-wandering.
references/model-catalog.md:89来自说明文档打开原文件
|---|---|---|| Virality Predictor (`brain_activity`) | Higgsfield | **Objective attention proxy for video creative testing.** Scores how effectively a clip captures and sustains attention, useful for hook validation, virality potential, ad review, and product/content focus. Takes a video input and returns a text report with overall score, peak second, sustain, and an Open report link. Raw `.glb` / `.bin` render artifacts stay in JSON/debug output. |

Skill 逻辑拆解

8 个说明模块

该 Skill 通过具有 Bash 权限的第三方 `higgsfield` CLI 提交图像、视频、3D 和音频生成任务，并默认等待远程任务完成。

查看原文
SKILL.md:24来自说明文档打开原文件
argument-hint: "[prompt-or-analysis-request] [--model <name>] [--image|--video <path-or-id>]"allowed-tools: Bash---
SKILL.md:29来自说明文档打开原文件
Submit jobs to any Higgsfield model. Wraps the `higgsfield` CLI. Covers generic image/video/3D/audio generation, Marketing Studio (branded ads, avatars, products, hooks, settings), and, secondarily, Virality Predictor video scoring.
SKILL.md:124来自说明文档打开原文件
3. **Validate quickly.** If unsure of params, run `higgsfield model get <jst> --json` once and pass only what's needed. Validate the preferred model before falling back to an older one. Use schema defaults otherwise. The server returns `adjustments` for non-fatal coercions (e.g. `aspect_ratio=99:99` → closest match) and a structured error for invalid declared-param values.4. **Submit and wait in one shot.** `higgsfield generate create <jst> [--prompt "..."] [media flags] [param flags] --wait`. Blocks until terminal status and prints the result on stdout. Tunables: `--wait-timeout 20m` (default 10m), `--wait-interval 5s` (default 3s). Virality Predictor does not need a prompt; pass `--video`.5. **Deliver.** For generated media and 3D assets, send the primary result URL plus a one-line summary (model, duration if video; GLB/asset URL for 3D). For Virality Predictor, deliver the scores, business interpretation, and the Open report link. Do not surface Virality Predictor `.glb`, `.bin`, or region-table internals in normal chat output.

本地媒体路径会由 CLI 自动上传；支持的内容包括照片、视频和音频，也包括人脸头像、声音参考及待分析的广告视频。

查看原文
references/media-inputs.md:7来自说明文档打开原文件
Each media flag accepts either a local file path or a UUID. The CLI auto-uploads paths before submission and auto-detects whether a UUID is an upload id (from `higgsfield upload create`) or a previous job id.
references/marketing-avatars.md:21来自说明文档打开原文件
## Creating a custom avatar```bashID=$(higgsfield upload create founder.png)URL=$(higgsfield upload create founder.png --json | jq -r .url)   # if you need cloudfront URLhiggsfield marketing-studio avatars create --name "Founder" --image $ID --image-url $URL```
references/media-inputs.md:103来自说明文档打开原文件
higgsfield generate create seed_audio \  --prompt "same voice, calmer delivery" \  --audio-references ./voice.wav \  --wait```

Marketing Studio 可在用户账户中创建并复用产品、头像、广告参考和品牌资料；部分对象会由后端异步处理和存储。

查看原文
SKILL.md:174来自说明文档打开原文件
- **Avatar** — presenter face. Curated `preset` (browse `higgsfield marketing-studio avatars list`) or `custom` (uploaded photos via `higgsfield marketing-studio avatars create`). For UGC modes, an avatar is optional if the brief clearly mentions a person; the backend can create a Soul Character automatically. Pass an avatar when the user wants a specific presenter.- **Product** — brand item with title + reference images. Imported from URL (`higgsfield marketing-studio products fetch --url ...`) or created from uploaded images (`higgsfield marketing-studio products create`).- **Webproduct** — App Store / web page version. Auto-routes when fetching App Store URLs.- **Hook** — reusable opening angle / ad hook. Browse with `higgsfield marketing-studio hooks list`. Hook text is prepended to the user's prompt; it does not replace `--prompt`.- **Setting** — reusable environment / scene context. Browse with `higgsfield marketing-studio settings list`.- **Ad reference** — reusable inspiration video that can be bound to an avatar and/or product. Created from an uploaded video (`--video-input <upload_id>`) or a previous generation job (`--job <job_id>`). Browse with `higgsfield marketing-studio ad-references list`. See `references/marketing-ad-references.md`.- **Brand kit** — captures a brand's identity (name, logo, hero images, colours, fonts, tone) for reuse across image generations. Created by handing in a website URL (`higgsfield marketing-studio brand-kits fetch --url https://… --wait`). See `references/marketing-brand-kits.md`.- **Ad format** — presets that drives the visual structure of a generated image (`headline`, `bullet-points`, etc.). Read-only, browse with `higgsfield marketing-studio ad-formats list`. Required input for `dtc-ads generate`.
references/marketing-ad-references.md:52来自说明文档打开原文件
The backend kicks off processing asynchronously. Newly created references start in `status: queued` then move to `in_progress` and finally `completed` (or `failed`).

Virality Predictor 会把已完成的视频发送给远程模型，并返回注意力、留存和传播潜力等评分及报告链接，供创意测试使用。

查看原文
SKILL.md:278来自说明文档打开原文件
Use Virality Predictor (`brain_activity`) when the user wants to evaluate a finished video as a business creative: hook strength, virality potential, attention, retention, or how well the content/product holds focus and minimizes distraction. Treat "Virality Predictor" as the customer-facing feature name; `brain_activity` is only the CLI/job_set_type.
SKILL.md:284来自说明文档打开原文件
The result is text, not a generated image/video. Report the overall score, peak hook second, sustain score, strongest/weakest regions, and report URL if present. Interpret it as an objective attention proxy for creative testing: higher Visual/Auditory/Language/Attention scores suggest stronger stimulus and focus; lower Default Mode is better because it suggests less mind-wandering.The CLI prints an Open report URL like `https://<app-domain>/apps/virality-predictor?resultJobId=<job_id>`. Send that URL for the visual report. Raw artifact URLs such as `brain_example_url`, `vertexMapBinaryUrl`, and `vertexMapUrl` are implementation details; mention them only when the user asks for raw data or implementation details.
从这里开始 · 工作说明SKILL.md
higgsfield-generate
连线表示工作说明包含的模块，不是实际运行顺序。点击模块可查看原文。 另有 2 个章节，可在原文件中查看。

文件引用关系图

12 处引用
哪些文件发起引用引用了什么
连线表示真实的文件引用，不是运行顺序。点击节点可高亮相关连线，并查看具体文件和原文位置。虚线表示还有文件需要定位。
文件与检查记录13 个文件

检查范围与遗漏

逐文件查看涉及的内容

下方列出本次涉及的原文范围；纳入检查不代表已查清所有问题。

  • SKILL.md已纳入全文
  • references/marketing-ad-references.md已纳入全文
  • references/marketing-avatars.md已纳入全文
  • references/marketing-brand-kits.md已纳入全文
  • references/marketing-dtc-ads.md已纳入全文
  • references/marketing-modes.md已纳入全文
  • references/marketing-products.md已纳入全文
  • references/marketing-setup-items.md已纳入全文
  • references/media-inputs.md已纳入全文
  • references/model-catalog.md已纳入全文
  • references/prompt-engineering.md已纳入全文
  • references/troubleshooting.md已纳入全文
  • references/workflows.md已纳入全文

这份报告只针对上方版本。我们看了拿到的代码和说明文件，没有实际运行 Skill，也没有检查它另外安装的软件包。因此，这不是“保证安全”的承诺；换了版本或使用环境，结果也可能不同。

  • SKILL.md工作说明
  • references/marketing-ad-references.md配套文件
  • references/marketing-avatars.md配套文件
  • references/marketing-brand-kits.md配套文件
  • references/marketing-dtc-ads.md配套文件
  • references/marketing-modes.md配套文件
  • references/marketing-products.md配套文件
  • references/marketing-setup-items.md配套文件
  • references/media-inputs.md配套文件
  • references/model-catalog.md配套文件
  • references/prompt-engineering.md配套文件
  • references/troubleshooting.md配套文件
  • references/workflows.md配套文件

代码和说明中提到的操作

运行命令
SKILL.md:24来自说明文档打开原文件
argument-hint: "[prompt-or-analysis-request] [--model <name>] [--image|--video <path-or-id>]"allowed-tools: Bash---
SKILL.md:36来自说明文档打开原文件
1. If `higgsfield` is not on `$PATH`, install it:   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh
SKILL.md:149来自说明文档打开原文件
```bashhiggsfield generate create gpt_image_2 --prompt "neon city at dusk" --aspect_ratio 16:9 --resolution 2k --wait
连接外部网站
SKILL.md:37来自说明文档打开原文件
   ```bash   curl -fsSL https://raw.githubusercontent.com/higgsfield-ai/cli/main/install.sh | sh   ```
SKILL.md:180来自说明文档打开原文件
- **Ad reference** — reusable inspiration video that can be bound to an avatar and/or product. Created from an uploaded video (`--video-input <upload_id>`) or a previous generation job (`--job <job_id>`). Browse with `higgsfield marketing-s …- **Brand kit** — captures a brand's identity (name, logo, hero images, colours, fonts, tone) for reuse across image generations. Created by handing in a website URL (`higgsfield marketing-studio brand-kits fetch --url https://… --wait`). See `references/marketing-brand-kits.md`.- **Ad format** — presets that drives the visual structure of a generated image (`headline`, `bullet-points`, etc.). Read-only, browse with `higgsfield marketing-studio ad-formats list`. Required input for `dtc-ads generate`.
SKILL.md:251来自说明文档打开原文件
# 1. Trigger fetch (returns the product id, import runs in the background)higgsfield marketing-studio products fetch --url https://shop.example.com/sneakers --wait
读取密钥或账号配置
references/troubleshooting.md:6来自说明文档打开原文件
- `Session expired.` → `higgsfield auth login`- `Stored credentials are for ... but current environment ...` → `higgsfield auth login` for the current API URL.- `Not authenticated.` → first `higgsfield auth login`.
读取了多少行
1,332
文件校验值（用于核对版本）
e4e312e2988a5ed185c718ba19c6595ee4579ddfd4d46dcd6d192f35de3e09d9