Claude-skill-registry-data dcode:mine-patterns

Analyze a work session to identify repeatable patterns that could become reusable skills or commands. Use at the end of a productive session to capture workflows worth automating, when noticing repetitive multi-step tasks, or when wanting to improve personal productivity with Claude Code.

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry-data
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry-data "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/mine-patterns" ~/.claude/skills/majiayu000-claude-skill-registry-data-dcode-mine-patterns && rm -rf "$T"
manifest: data/mine-patterns/SKILL.md
source content

Mine Patterns

Turn productive sessions into reusable skills.

For designers who think: "I keep doing this same thing... there must be a better way."

Why This Matters

Every time you solve a problem, you're creating a workflow. Most workflows get forgotten. This skill helps you capture the good ones before they disappear.

Instructions

1. Review Session Activities

Look at what was accomplished:

  • What tasks were repeated or could be repeated?
  • What multi-step workflows were performed?
  • What required specific domain knowledge?
  • What felt tedious or error-prone?

2. Identify Skill Candidates

Good skills have these traits:

TraitWhy It Matters
RepeatableWill be done again in future sessions
Multi-stepMore than a single action
GeneralizableWorks across different contexts
Time-savingAutomates tedious or error-prone work
Knowledge-heavyRequires remembering specific patterns

3. Categorize by Value

High value - Build these first:

  • Complex workflows done frequently
  • Tasks where mistakes are costly
  • Processes that require specific conventions

Medium value - Build when you have time:

  • Useful but less frequent tasks
  • Nice-to-have automations

Lower value - Maybe don't bother:

  • One-off investigations
  • Highly context-specific tasks

4. Present Suggestions

For each potential skill, document:

## Suggested Skill: {name}

**Problem it solves:** {What pain point does this address?}

**Trigger:** {When would someone invoke this?}

**Steps it automates:**
1. {Step 1}
2. {Step 2}
3. {Step 3}

**Value:** High / Medium / Low

**Complexity to build:** Quick (1-2 hrs) / Medium (half-day) / Complex (day+)

5. Help Build the Chosen Skills

If the user wants to create a skill:

  1. Draft the SKILL.md with proper frontmatter:
---
name: skill-name
description: Clear description of what it does and when to use it
---
  1. Write clear instructions
  2. Include examples
  3. Test it on a real task

Example Output

Based on this session, here are potential skills:

High Value

1. design-token-audit

  • Problem: Finding inconsistent colors/spacing across a codebase
  • Trigger: "Audit this component for design system compliance"
  • Steps: Scan for hardcoded values, compare against tokens, report violations
  • Complexity: Medium

2. responsive-check

  • Problem: Verifying components work at all breakpoints
  • Trigger: Before PR, after styling changes
  • Steps: Identify breakpoints, list what changes at each, flag potential issues
  • Complexity: Quick

Medium Value

3. figma-to-code-notes

  • Problem: Translating design specs into implementation notes
  • Trigger: Starting implementation of a new design
  • Steps: Extract spacing, colors, typography, create implementation checklist
  • Complexity: Medium

Which of these would you like to create?

Meta Note

This skill is itself an example of workflow mining—it was created by noticing that "identifying reusable patterns" was a repeatable, valuable task.