Chat Search

Search and find relevant chat messages from Feishu or Telegram using semantic search with similarity scores and message source labels.

Security Scan

The skill claims to perform semantic search over Feishu and Telegram chats but does not explain how it will access those messages (no credential/ingest path), which is an important mismatch that should be clarified before use.

Purpose & Capability

The skill's name/description say it searches Feishu and Telegram chat history, but the instructions do not declare or request any means of accessing those platforms (no API tokens, export/upload steps, or connectors). Searching remote chat systems normally requires credentials or a connector — their absence is an incoherence between claimed purpose and required inputs.

Instruction Scope

SKILL.md instructs use of a local Qdrant instance and FastEmbed for embeddings and lists triggers, but it does not describe how chat data is obtained, indexed, or whether the agent will request uploads or credentials. The instructions also tell the operator to run docker and pip commands, but are otherwise vague about the runtime data flows, which grants the agent broad unspecified discretion.

Install Mechanism

No formal install spec in the registry, but SKILL.md recommends running 'docker run qdrant/qdrant' and 'pip install fastembed' and requires Node.js. These are standard/traceable sources (Docker Hub, pip) and not an unusual install pattern for a vector-search skill.

Credentials

The registry metadata lists no required env vars or credentials, yet the skill's function (searching Feishu/Telegram) normally requires platform tokens or access to exported chat data. The omission is disproportionate: either the skill expects the agent to ask the user for credentials or uploads at runtime, or the author left out critical details.

Persistence & Privilege

always is false and there is no install-time code or persistent config declared in the registry. The skill does not request elevated or always-on privileges in the manifest.

What to consider before installing

This skill claims it can search your Feishu and Telegram chats but the instructions do not say how it will get those messages (no API tokens, connectors, or upload steps). Before installing or using it, ask the author: (1) how are chat messages ingested and indexed — do you need to provide exports or API credentials? (2) does embedding and search run entirely locally (Qdrant + FastEmbed) or will anything be sent to remote services? (3) what exact commands or code the agent will run when triggered? Also be cautious about giving any chat platform tokens or uploading private chat history until you confirm the data flow and where embeddings/search indexes are stored. The install commands shown (docker run qdrant, pip install fastembed) are standard, but because this is an instruction-only skill with no code reviewed, get more detail about runtime behavior and data access before proceeding.

Chat Search Skill

飞书/Telegram 聊天记录语义搜索

功能

• 语义搜索聊天记录
• 支持中文分词
• 显示相似度分数
• 区分消息来源（飞书/Telegram）
• 使用 Qdrant 向量数据库
• 使用 FastEmbed 生成中文向量

触发方式

当用户说以下内容时自动触发：

• "搜索XXX" / "搜素XXX" / "查找XXX"
• "帮我搜一下XXX"
• "找一下关于XXX的对话"

依赖

• Qdrant 向量数据库 (localhost:6333)
• Python FastEmbed (BGE 中文模型)
• Node.js

使用示例

用户：帮我搜索 Docker 相关的对话
小T：→ 执行搜索

🔍 搜索 "Docker" 的结果：

1. 🐦 [feishu] 小T (2026-03-20 09:17)
   Docker 安装成功了
   相似度: 58.2%

安装

需要先安装依赖：

# Qdrant
docker run -d --name qdrant -p 6333:6333 qdrant/qdrant

# Python FastEmbed
pip install fastembed