dbskill AI Skills Toolbox

repository·main·Indexed 27 days ago

https://github.com/dontbesilent2025/dbskill

A Chinese AI Skills toolbox for entrepreneurs and content creators featuring 29 specialized skills for business diagnosis, content creation, and knowledge management. It includes a smart entry point (/dbs) for automated skill selection and supports Agents like Claude Code, Doubao, WorkBuddy, and Codex. The system provides tools for managing local knowledge bases, an Atomic Knowledge Base schema for RAG construction, and a Content Structured System for combinatorial content creation using Node.js scripts and Obsidian linking.

Tokens
103.1K
Snippets
116
Records
599
Agent score
94%

What's inside dbskill

  1. Understand the collection structure of dontbesilent Open Tweet Collection

    main

    The dontbesilent Open Tweet Collection (from 2024 to present) is a curated repository of insights organized by date. Each entry typically includes:

    • Type: Categorized as 主贴 (Main Post) or 引用 (Quote).
    • Content: The core text or quoted material.
    • Metadata:
      • Original Post Link: A direct URL to the source on X (formerly Twitter).
      • Theme (主题): The high-level subject matter (e.g., AI & Tools, Business & Product).
      • Expression Mode (表达): The nature of the insight (e.g., Method, Viewpoint, Observation, Problem).
      • Tags (标签): Specific keywords for filtering (e.g., AI, Philosophy, Content Production).
  2. Understand dbs operating modes

    main

    The dbs toolkit operates in three distinct modes based on your current context and input:

    1. Mode C (Onboarding Tutorial): Triggered by /dbs 新手入门 or explicit requests for guidance. It explains how the system works and helps you complete your first actual task.
    2. Mode A (Pre-task Routing): Triggered when you have a clear need but haven't used a specific skill yet. dbs identifies your requirement and routes you to the correct specialized Skill.
    3. Mode B (Post-task Navigation): Triggered after a dbs-* skill has produced an output (e.g., a report, checklist, or analysis). dbs reads the conclusion of the previous task and suggests the most valuable next direction.
  3. Understand the content collection structure and rules

    main

    The dontbesilent-开源推文集 is a collection of curated tweets (from 2024 to present) categorized by date, type (e.g., 主贴 for main posts, 引用 for quotes), and metadata. Each entry typically includes:

    • Content: The core insight or text.
    • Original Post Link: A link to the source on X (formerly Twitter).
    • Theme: The high-level topic (e.g., 商业与产品 - Business & Product).
    • Expression Type: The mode of expression (e.g., 案例 - Case Study, 观点 - Opinion, 方法 - Method).
    • Tags: Keywords for searching (e.g., 创业, 内容创作).
  4. Insights on Benchmarking and Imitation

    main

    The benchmark_对标方法论.md knowledge pack contains various insights regarding benchmarking (对标), imitation (模仿), and competitive strategy. Key themes include:

    • Imitation vs. Innovation: Distinguishing between imitating product appearance (e.g., car aesthetics) versus imitating brand culture or core innovation.
    • Competitive Strategy: Recognizing that the strongest competitors often play a different game (e.g., focusing on high customer lifetime value vs. low acquisition cost).
    • Market Validation: Using existing products as benchmarks because they represent a validated market, while noting the difficulty in replicating original brand culture.
    • Benchmarking Layers: The importance of benchmarking "originality" (原创性) as a critical layer of analysis.
    • Content and Marketing: Strategies for competing in markets where product differentiation is impossible (e.g., identical supply chains) by focusing entirely on content and marketing differentiation.
  5. Understand the content structure of the dontbesilent Open Tweet Collection

    main

    The dontbesilent-开源推文集 is a curated collection of tweets (from 2024 to present) organized by date. Each entry typically includes:

    • Date: The timestamp of the post (e.g., 2025-12-18).
    • Type: Categorized as 引用 (Quote) or 主贴 (Main Post).
    • Content: The actual text of the tweet.
    • Metadata:
      • 原帖: A direct link to the original X (formerly Twitter) post.
      • 主题: The core subject matter (e.g., 认知与语言, 内容与传播).
      • 标签: Relevant tags for categorization (e.g., 赚钱, 商业模式, 认知).

    This structure allows users to navigate the collection by date, topic, or specific themes like content production, business models, and cognitive science.

  6. Browse the dontbesilent Open Source Tweet Collection

    main
    The dontbesilent open-source tweet collection (from 2024 to present) serves as a repository of insights, business philosophies, and cognitive frameworks. Users can browse entries categorized by date, type (e.g., 主贴 for main posts, 引用 for quotes), theme (e.g., 商业与产品 - Business & Product, 认知与语言 - Cognition & Language), and tags (e.g., 创业 - Entrepreneurship, AI). Each entry typically includes the core insight, the original X (formerly Twitter) link, the theme, and relevant tags.
  7. Understand the structure of the dontbesilent Open Tweet Collection

    main

    The dontbesilent Open Tweet Collection is a curated repository of social media posts (primarily from X/Twitter) spanning from 2024 to the present. Each entry is categorized by date, type (e.g., 主贴 for main posts, 引用 for quotes), and includes metadata for thematic organization.

    Each record typically contains:

    • Content: The text of the tweet.
    • Original Link: A link to the source post on X.
    • Theme (主题): The high-level domain (e.g., 商业与产品 - Business & Product, 行动与心理 - Action & Psychology).
    • Expression Type (表达): The nature of the content (e.g., 观点 - Opinion, 观察 - Observation, 经验 - Experience).
    • Tags (标签): Specific keywords for granular searching (e.g., 商业模式 - Business Model, 用户需求 - User Needs).

    This collection serves as a knowledge base for insights on business, product design, psychology, and AI tools.

  8. Understand the dbs-script-flow check dimensions

    main

    The skill evaluates scripts across three specific dimensions to ensure high viewer retention:

    1. Logical Connection (Between paragraphs): Checks if there is a logical leap between the end of one paragraph and the start of the next. It looks for missing transitions or unclosed topics.
    2. Information Density (Within paragraphs): Checks for redundancy, circular reasoning, or "correct nonsense" (filler content) that doesn't add value.
    3. Oral Fluency (Sentence level): Checks if sentences are too long (over 30 characters), too formal/written, or contain unexplained jargon that makes them difficult to speak or hear naturally.
  9. Use the dbs-knowledge skill for folder-based knowledge bases

    main

    The dbs-knowledge skill allows AI agents to transform existing local folders into stable, manageable knowledge bases. It handles the creation, population, querying, organization, auditing, and slimming of local directories without requiring external databases, vector stores, or RAG systems.

    Core Capabilities:

    • Build (建): Create new folder-based knowledge bases or add navigation to existing ones.
    • Store (存): Place new materials in appropriate locations, preserve sources, and manage duplicate names and version relationships.
    • Use (用): Locate original files from the knowledge base to provide comprehensive answers based on evidence.
    • Check (查): Audit for broken paths, version conflicts, scattered files, duplicate data, and maintenance gaps.

    Key Concept: When discussing technical implementation files like SOURCE_OF_TRUTH.md, refer to them as the "Knowledge Base Navigation" (知识库导航) to the user.

  10. Browse the dontbesilent Open Tweet Collection

    main

    The dontbesilent Open Tweet Collection (from 2024 to present) is a curated repository of insights, observations, and methods shared by dontbesilent. The collection is organized by date and categorized by themes such as:

    • Business & Product (商业与产品): Insights on business logic, market demand, and product innovation.
    • Content & Communication (内容与传播): Strategies for personal IP, traffic acquisition, content monetization, and platform utilization.
    • Cognition & Language (认知与语言): Reflections on questioning abilities and information retrieval.

    Each entry typically includes the original text, a link to the original X (formerly Twitter) post, a theme, an expression type (e.g., Observation, Method, Case, Viewpoint), and relevant tags.

  11. Insights on Execution and Psychology (Action Framework)

    main

    This segment of the action_心理诊断框架 (Psychological Diagnosis Framework) provides a collection of insights regarding the relationship between psychology, execution, and business success. Key themes include:

    • Execution vs. Knowledge: Execution is a skill that generates knowledge through practice. Those who wait for perfect knowledge before acting often fail to start.
    • Psychology as a Business Barrier: Many perceived business or technical problems (e.g., procrastination, fear of failure, lack of consistency) are actually underlying psychological issues.
    • Entrepreneurship and Emotional Regulation: High levels of internal friction (内耗) and low self-esteem are significant barriers to entrepreneurship. Success often requires the ability to handle failure without excessive emotional distress.
    • AI and Productivity: Using AI (like Claude) for psychological support or to automate tasks can help bridge the gap between intention and execution, but the tool itself cannot solve fundamental psychological resistance.
    • The Value of Action: The primary value of execution is that completing a task provides the necessary feedback and information to understand the next steps.