second-brain-skills

🥈Silver

A collection of Claude Skills that transform Claude Code into a second brain for knowledge work. These skills enable progressive disclosure of context, making Claude hyper-capable for specific tasks. Ideal for operations teams looking to automate and streamline knowledge-intensive processes.

682540Updated 2w ago

Intermediate30min to implementautomation

Saves ~360 min per use

Quick InstallView Source

git clone https://github.com/coleam00/second-brain-skills.git

Works with:

Claude

Overview

About This Skill

How to Use

[{"step":"Define the core task and context","description":"Start by identifying the specific knowledge work task you want to automate (e.g., lead qualification, bug triage, contract review). Gather the initial context or data required for the skill to operate. For example, if building a lead qualification skill, collect your CRM data, ICP criteria, and historical conversion data.","tip":"Use Claude Code to pre-process your data (e.g., clean CSV files, validate fields) before feeding it into the skill. This ensures the skill starts with high-quality input."},{"step":"Design the progressive disclosure flow","description":"Break the task into 3–5 subtasks that progressively reveal context. For each subtask, define: (1) the input required, (2) the processing logic, and (3) the output format. For example, a bug triage skill might have subtasks for data ingestion, severity classification, root cause analysis, and resolution suggestions.","tip":"Use a flowchart tool like Mermaid.js to visualize the skill’s flow before coding. This helps identify gaps or redundant steps early."},{"step":"Implement the skill in Claude Code","description":"Write the Claude Skill using Claude Code, leveraging tools like `pandas` for data manipulation, `scikit-learn` for scoring models, or `langchain` for dynamic context disclosure. Include placeholders for user inputs (e.g., [CRITERIA], [RESOURCES]) to make the skill adaptable.","tip":"Start with a minimal viable version (e.g., just the data ingestion and first subtask) and iterate. Use Claude’s inline code execution to test each subtask in real-time."},{"step":"Test and refine with real-world data","description":"Run the skill with a small batch of real data to validate its outputs. Compare the skill’s results with manual processes to identify discrepancies. For example, if the skill misclassifies 10% of leads, adjust the qualification rules or scoring weights.","tip":"Use Claude’s `diff` tool to compare the skill’s outputs with your expected results. This helps pinpoint where the skill is deviating from your goals."},{"step":"Deploy and iterate","description":"Deploy the skill in your workflow (e.g., as a Claude Skill in your CRM or as a standalone tool). Monitor its performance and gather feedback from users. Use this data to continuously improve the skill, such as adding new subtasks or refining the scoring logic.","tip":"Set up a feedback loop where users can rate the skill’s outputs (e.g., \"Was this lead qualification helpful? Yes/No\"). This data can be fed back into the skill to improve future iterations."}]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/coleam00/second-brain-skills

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Transform this [TASK_DESCRIPTION] into a Claude Skill that acts as a second brain for knowledge work. Design it to progressively disclose context, break the task into subtasks, and provide actionable outputs at each stage. Include [CONTEXT] for initial setup and [USER_INPUT] for dynamic adaptation. Example: 'Create a Claude Skill that helps a customer support agent resolve [ISSUE_TYPE] tickets. Start with a triage step that classifies the ticket using [CRITERIA], then provide a step-by-step resolution guide with [RESOURCES], and finally suggest follow-up actions for the agent.'

Example Output

Here’s a Claude Skill designed to help a sales operations team automate lead qualification for a SaaS company. The skill progressively discloses context by first analyzing raw lead data, then applying qualification rules, and finally generating a prioritized action plan for the sales team.

**Step 1: Data Ingestion and Initial Context**
The skill starts by ingesting a CSV file of 5,000 leads from [CRM_NAME] with fields like `lead_id`, `company_name`, `industry`, `employee_count`, `revenue`, `last_touch_date`, and `source`. It calculates basic statistics: 60% of leads are from LinkedIn, 30% from webinars, and 10% from referrals. The average employee count is 250, with a median revenue of $12M. The skill flags that 1,200 leads haven’t been contacted in over 90 days and are at risk of churning.

**Step 2: Qualification Rules Application**
Next, the skill applies a multi-layered qualification framework. First, it filters leads based on ICP (Ideal Customer Profile) criteria: employee count between 100–1,000, revenue >$5M, and industry in [TECH, FINANCE, HEALTHCARE]. This reduces the pool to 1,800 leads. Then, it scores each lead using a custom algorithm that weights:
- Engagement score (30%): based on email opens, website visits, and demo requests in the last 30 days.
- Firmographic fit (40%): alignment with ICP.
- Behavioral signals (30%): such as attending a product webinar or downloading a whitepaper.
The top 200 leads are scored >85/100, with Company X (employee count: 450, revenue: $22M) scoring the highest at 94/100 due to high engagement and perfect ICP fit.

**Step 3: Actionable Outputs and Next Steps**
The skill generates a prioritized action plan for the sales team:
1. **Immediate Outreach**: 50 leads with scores >90 should be contacted within 48 hours. The skill provides a templated email sequence and suggests calling during peak hours (10 AM–12 PM EST).
2. **Nurture Campaign**: 300 leads scoring 70–89 should enter a 3-email nurture sequence over 2 weeks, with a focus on case studies relevant to their industry.
3. **Re-engagement**: For the 1,200 stale leads, the skill drafts a win-back campaign with a special offer (e.g., 20% discount on first-year subscription) and recommends a phone call script.
4. **Data Hygiene**: The skill identifies 50 leads with invalid email addresses and flags them for CRM cleanup.

**Step 4: Progressive Disclosure for Agents**
The skill is designed to reveal only the most relevant context at each stage. For example, when an agent clicks on a lead, the skill shows:
- A summary card with the lead’s score, ICP fit, and last interaction date.
- A "Why This Lead Matched" section explaining the qualification criteria.
- A "Next Best Action" button that generates a personalized email draft or call script based on the lead’s behavior.

**Step 5: Continuous Learning Loop**
The skill includes a feedback mechanism where agents can mark leads as "qualified" or "unqualified" after outreach. This data is fed back into the qualification algorithm to improve future scoring. For example, after 100 interactions, the skill notices that leads from webinars are 20% more likely to convert than those from LinkedIn, so it adjusts the scoring weights accordingly.

By the end of the process, the sales team has a hyper-focused list of 50 high-priority leads, a clear nurture strategy for the rest, and actionable steps to improve data quality—all while the skill adapts based on real-time feedback.