Signus Font Signature
User-facing language
- - Always write user-facing text in English unless the user explicitly asks for another language.
Runtime requirements
- - Node.js 18+.
- Install dependencies in this skill folder before first run:
CODEBLOCK0
Inputs
Provide exactly one identity source:
- -
name (e.g., "Allon Mason"), OR - INLINECODE2 +
lastName, OR - INLINECODE4 (e.g.,
"AM")
Optional:
- -
count — max number of images to return.
Validation
- - Accept exactly one identity source.
- Reject empty/blank values.
- If
count is provided, treat as numeric limit.
Run
CODEBLOCK1
Examples:
CODEBLOCK2
CODEBLOCK3
Detailed execution flow
- 1. Parse payload and normalize identity into
name. - Create output directory:
-
~/.openclaw/media/signatures-font/<name>-<timestamp>/
- 3. Request font generation using fixed trusted host (
https://api.signus.ai) and endpoint order:
- primary:
POST /api/signus/v0/signature-generations/font
- fallback:
POST /api/signus/v0/users/me/signature-generations/font
- 4. Process response by content-type:
- ZIP/octet-stream: extract with in-process JS unzip library (
adm-zip), then collect image files.
- JSON: read
payload.thisPageItems[], then download each image from:
-
/api/signus/v0/signature-generations/font/{generationId}/signatures/{generatedSignatureId}/{clean|watermark}.png
- 5. Return:
-
count
-
directory
-
signatures[]: { id, filePath }
- 6. Send images to chat via
message action=send and media=<filePath>.
Authentication model
- - This skill does not handle API tokens or env-based credentials.
- It assumes the primary public endpoint can be used without explicit
Authorization header. - The
/users/me/... fallback may work only where implicit session/auth exists. - If your deployment requires explicit auth, update the script design first (do not inject secrets into payloads by default).
Security notes
- - No environment-variable reads.
- No shell command execution.
- Network target is fixed to
https://api.signus.ai. - Writes only under
~/.openclaw/media/signatures-font/. - Keep this skill separate from
signus-signature (AI websocket handwritten signatures).
Signus 字体签名
面向用户的语言
- - 除非用户明确要求使用其他语言,否则始终使用英文编写面向用户的文本。
运行时要求
- - Node.js 18+。
- 首次运行前,请在此技能文件夹中安装依赖:
bash
cd /home/node/.openclaw/workspace/skills/signus-font-signature && npm install
输入
提供且仅提供一个身份来源:
- - name(例如 Allon Mason),或
- firstName + lastName,或
- initials(例如 AM)
可选参数:
验证规则
- - 仅接受一个身份来源。
- 拒绝空值/空白值。
- 如果提供了 count,则将其视为数字限制。
运行命令
bash
node /home/node/.openclaw/workspace/skills/signus-font-signature/scripts/generate
fontsignatures.js
示例:
bash
node /home/node/.openclaw/workspace/skills/signus-font-signature/scripts/generatefontsignatures.js {name:Allon Mason,count:10}
bash
node /home/node/.openclaw/workspace/skills/signus-font-signature/scripts/generatefontsignatures.js {firstName:Allon,lastName:Mason}
详细执行流程
- 1. 解析负载并将身份信息标准化为 name。
- 创建输出目录:
- ~/.openclaw/media/signatures-font/-/
- 3. 使用固定的可信主机(https://api.signus.ai)请求字体生成,端点顺序如下:
- 主端点:POST /api/signus/v0/signature-generations/font
- 备用端点:POST /api/signus/v0/users/me/signature-generations/font
- 4. 根据内容类型处理响应:
- ZIP/octet-stream:使用进程内 JS 解压库(adm-zip)解压,然后收集图片文件。
- JSON:读取 payload.thisPageItems[],然后从以下路径下载每张图片:
- /api/signus/v0/signature-generations/font/{generationId}/signatures/{generatedSignatureId}/{clean|watermark}.png
- 5. 返回:
- count
- directory
- signatures[]: { id, filePath }
- 6. 通过 message action=send 和 media= 将图片发送到聊天界面。
认证模型
- - 此技能不处理 API 令牌或基于环境的凭据。
- 它假设无需显式的 Authorization 头即可使用主公共端点。
- /users/me/... 备用端点可能仅在存在隐式会话/认证的情况下有效。
- 如果您的部署需要显式认证,请先更新脚本设计(默认情况下不要将密钥注入负载)。
安全说明
- - 不读取环境变量。
- 不执行 shell 命令。
- 网络目标固定为 https://api.signus.ai。
- 仅写入 ~/.openclaw/media/signatures-font/ 目录。
- 保持此技能与 signus-signature(AI WebSocket 手写签名)分离。