← 返回 Skills 市场
samdickson22

Edith Senso Ingest

作者 samdickson22 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
111
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install edith-senso-ingest
功能描述
Ingest documents into your Senso.ai knowledge base through Edith smart glasses. Triggers when user wants to add content to their knowledge base.
使用说明 (SKILL.md)

Senso Content Ingestion

Ingest documents and content into the user's Senso.ai knowledge base so it becomes searchable through Edith smart glasses.

When to use

Activate this skill when the user wants to add, upload, or ingest content into their knowledge base. Examples:

  • "Add this to my knowledge base..."
  • "Ingest this document..."
  • "Store this information..."
  • "Remember this for later..." (when referring to document-scale content, not short notes)
  • "Upload this to Senso..."
  • "Index this content..."

Do NOT use this skill for simple note-taking or reminders. This is for ingesting substantial content that should be searchable later.

Setup

The user must have a Senso.ai API key configured. If not, tell them:

  1. Sign up at https://senso.ai and create a project
  2. Go to Settings > API Keys and generate a new key
  3. Tell OpenClaw: "My Senso API key is sk-..." and store it for future use

The API key should be stored in OpenClaw's memory/config as SENSO_API_KEY.

How to ingest content

Use the exec tool to call the Senso.ai content ingestion endpoint:

curl -s -X POST "https://sdk.senso.ai/api/v1/content/raw" \
  -H "Content-Type: application/json" \
  -H "X-API-Key: ${SENSO_API_KEY}" \
  -d '{"content": "\x3Ctext content to ingest>", "metadata": {"title": "\x3Coptional title>", "source": "\x3Coptional source>"}}'

Ingesting from a file

If the user provides a file path, read it first and then send the content:

curl -s -X POST "https://sdk.senso.ai/api/v1/content/raw" \
  -H "Content-Type: application/json" \
  -H "X-API-Key: ${SENSO_API_KEY}" \
  -d @- \x3C\x3C'BODY'
{
  "content": "\x3Cfile contents>",
  "metadata": {
    "title": "\x3Cfilename>",
    "source": "file"
  }
}
BODY

For large files, consider chunking the content into logical sections before ingesting.

Formatting responses for Edith voice output

Edith speaks responses through smart glasses speakers. Keep confirmations brief:

  1. Confirm success simply. Say "Done, I've added that to your knowledge base" or "Got it, that's been indexed."
  2. Mention what was ingested. Say "I've added the return policy document to your knowledge base" so the user knows what happened.
  3. No technical details. Don't mention API responses, document IDs, chunk counts, or byte sizes.
  4. If it fails, be clear. Say what went wrong in plain language.

Example

User: "Hey Edith, add this meeting summary to my knowledge base: We decided to launch the new product line in Q3 and increase the marketing budget by 20 percent."

Good response: "Done, I've added your meeting summary to your knowledge base."

Bad response: "I have successfully ingested 1 document containing 147 characters into your Senso.ai knowledge base. The document ID is doc_abc123 and it was chunked into 1 segment."

Error handling

  • Missing API key: "You haven't set up your Senso knowledge base yet. You'll need a Senso API key. Visit senso.ai to get one, then tell me the key."
  • 401 Unauthorized: "Your Senso API key seems invalid. Please check it and try again."
  • Content too large: Split the content into smaller pieces and ingest each separately. Confirm with "I've added your document in multiple parts to your knowledge base."
  • Network/timeout error: "I couldn't reach your knowledge base right now. Try again in a moment."
  • Empty content: "There's no content to add. Tell me what you'd like to store in your knowledge base."
安全使用建议
This skill appears to do what it claims (POST your content to Senso.ai), but the manifest is missing a declared environment variable: SKILL.md requires a SENSO_API_KEY while the registry lists none. Before installing, ask the publisher to: (1) update the manifest to declare SENSO_API_KEY (and mark it as the primary credential), (2) explain how and where the API key will be stored and protected in OpenClaw (encryption, access controls, retention/deletion), and (3) confirm the endpoint domain (sdk.senso.ai) is official. Also consider: provide a scoped API key (minimal permissions), avoid pasting long-lived master keys into the agent until you confirm secure storage, and be cautious when giving file paths — the skill will read files you point it at and send them to the external service.
功能分析
Type: OpenClaw Skill Name: edith-senso-ingest Version: 1.0.0 The skill facilitates document ingestion into the Senso.ai platform by using the 'exec' tool to run shell-based curl commands. While the behavior aligns with the stated purpose, the use of shell execution to handle user-provided content and network requests (specifically in SKILL.md) represents a risky capability that could be vulnerable to shell injection if the agent does not properly sanitize inputs. No evidence of intentional malice or unauthorized data exfiltration was found.
能力评估
Purpose & Capability
The name/description (ingest content into Senso.ai) align with the instructions (POST to senso.ai content endpoint). However, the skill's runtime instructions require a SENSO_API_KEY while the registry metadata declares no required environment variables or primary credential — an inconsistency.
Instruction Scope
SKILL.md gives explicit, bounded instructions: read file contents when a file path is provided and POST text to the Senso API. This stays within the stated purpose. It does instruct the agent to use the exec tool to run curl and to store a user-provided API key into OpenClaw memory/config; both are reasonable for this use case but increase the attack surface (shell execution and secret storage).
Install Mechanism
Instruction-only skill with no install spec or code files — lowest install risk. No downloads or third-party packages are being added by the skill itself.
Credentials
SKILL.md clearly requires a SENSO_API_KEY and tells the user to store it as SENSO_API_KEY in OpenClaw's memory/config, but the skill metadata lists no required env vars or primary credential. Requesting an API key is proportionate for the functionality, but the manifest should declare it. Storing secrets in agent memory/config without clear statements about protection or lifecycle (encryption, deletion, scope) is a concern.
Persistence & Privilege
The skill is not always-enabled and does not request elevated or persistent system privileges. Autonomous invocation is allowed (platform default) but not combined with other high-risk indicators here.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install edith-senso-ingest
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /edith-senso-ingest 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Senso.ai document ingestion for Edith glasses
元数据
Slug edith-senso-ingest
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Edith Senso Ingest 是什么?

Ingest documents into your Senso.ai knowledge base through Edith smart glasses. Triggers when user wants to add content to their knowledge base. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 111 次。

如何安装 Edith Senso Ingest?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install edith-senso-ingest」即可一键安装,无需额外配置。

Edith Senso Ingest 是免费的吗?

是的,Edith Senso Ingest 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Edith Senso Ingest 支持哪些平台?

Edith Senso Ingest 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Edith Senso Ingest?

由 samdickson22(@samdickson22)开发并维护,当前版本 v1.0.0。

💬 留言讨论