Skill Hunter

Use when: You need to discover existing skills from GitHub repositories that use AI agent frameworks (pi, cursor, opencode, etc.) - searches across multiple agent directories to find reusable skill patterns, implementations, and best practices.

Core Capability

Search GitHub for skills in AI agent repositories using path-based search to discover:

• Skill implementations from other projects
• Common patterns and approaches
• Reusable skill definitions
• Best practices from the community

Prerequisites

• gh CLI installed and authenticated
• jq for JSON processing
• $DOTFILES_REPO_ROOT environment variable set (or will auto-detect)
• Skills stored in $DOTFILES_REPO_ROOT/ai/files/skills/

GitHub Search Syntax

The core search pattern targets skill directories in AI agent repos:

bash
1path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/

This pattern finds files in paths like:

• .pi/skills/
• .ai/skills/
• .opencode/skill/
• pi/skills/
• agents/skills/
• cursor/skills/
• factory/skill/

Implementation Pattern

1. Basic Skill Search

Search for all skills matching a keyword:

bash
1# Search for skills related to "testing"
2gh search code "testing path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
3
4# Search for zellij skills
5gh search code "zellij path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20

2. Find Skills by File Type

Search for specific file types (e.g., all SKILL.md files):

bash
1# Find SKILL.md files
2gh search code "SKILL.md path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 50
3
4# Find skill implementations in TypeScript
5gh search code "extension:ts path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
6
7# Find shell script skills
8gh search code "extension:sh path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20

3. Language-Specific Skills

bash
1# Find Python skills
2gh search code "language:python path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
3
4# Find JavaScript/TypeScript skills
5gh search code "language:typescript path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
6
7# Find Rust skills
8gh search code "language:rust path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20

4. Find Skills by Domain

bash
1# Database skills
2gh search code "database OR postgres OR mysql path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
3
4# DevOps skills
5gh search code "docker OR kubernetes OR terraform path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
6
7# Web development skills
8gh search code "react OR vue OR angular path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20
9
10# Testing skills
11gh search code "playwright OR selenium OR jest path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 20

5. Discover Popular Repositories

Find repos with the most skills:

bash
1# Search and group by repository
2gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" --limit 100 --json repository,path | \
3  jq -r '.[] | .repository.fullName' | sort | uniq -c | sort -rn | head -10

6. Download and Inspect Skills

Once you find an interesting skill:

bash
1# View a specific file directly
2gh browse <owner>/<repo> --branch main <path-to-skill-file>
3
4# Clone the entire repository
5gh repo clone <owner>/<repo>
6
7# Download a specific file
8gh api repos/<owner>/<repo>/contents/<path-to-skill-file> \
9  --jq '.content' | base64 -d > downloaded-skill.md

Complete Workflow Example

Scenario: Finding and Adapting a Zellij Skill

bash
1# 1. Search for zellij skills
2echo "Searching for zellij skills..."
3gh search code "zellij path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
4  --limit 20 \
5  --json repository,path,url
6
7# 2. Search more broadly for terminal multiplexer skills
8echo "Searching for terminal/tmux/multiplexer skills..."
9gh search code "tmux OR terminal OR multiplexer path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
10  --limit 20 \
11  --json repository,path,url
12
13# 3. Examine a promising result
14REPO="username/repo-name"
15SKILL_PATH=".pi/skills/terminal-management/SKILL.md"
16
17gh api "repos/${REPO}/contents/${SKILL_PATH}" \
18  --jq '.content' | base64 -d | less
19
20# 4. Download for adaptation
21gh api "repos/${REPO}/contents/${SKILL_PATH}" \
22  --jq '.content' | base64 -d > /tmp/reference-skill.md
23
24echo "Downloaded to /tmp/reference-skill.md for review"

Advanced Search Patterns

Find Recently Updated Skills

bash
1# Skills updated in the last month
2gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/ pushed:>$(date -d '1 month ago' +%Y-%m-%d)" \
3  --limit 30

Find Skills by Star Count

bash
1# Skills from popular repositories (>100 stars)
2gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/ stars:>100" \
3  --limit 30

Find Skills with Specific Patterns

bash
1# Skills that use subagents
2gh search code "subagent path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 20
4
5# Skills that use MCP (Model Context Protocol)
6gh search code "MCP OR 'Model Context Protocol' path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
7  --limit 20
8
9# Skills with tool implementations
10gh search code "tool OR function_call path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
11  --limit 20

Output Processing

Extract and Format Results

bash
1# Get a clean list of repositories with skills
2gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 100 \
4  --json repository | \
5  jq -r '.[].repository | "\(.fullName) - \(.description // "No description")"' | \
6  sort -u
7
8# Create a markdown report of found skills
9gh search code "SKILL.md path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
10  --limit 50 \
11  --json repository,path,url | \
12  jq -r '.[] | "- [\(.repository.fullName)](\(.repository.url)) - [\(.path)](\(.url))"' > skills-report.md

Common Use Cases

1. Before Writing a New Skill

bash
1# Check if someone already solved this problem
2SKILL_TOPIC="docker"
3gh search code "${SKILL_TOPIC} path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
4  --limit 20 \
5  --json repository,path,url | \
6  jq -r '.[] | "[\(.repository.fullName)] \(.path)\n  \(.url)\n"'

2. Finding Best Practices

bash
1# Find skills with comprehensive documentation
2gh search code "## Prerequisites OR ## Requirements OR ## Installation path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 30

3. Discovering Skill Patterns

bash
1# Find skills that use specific frameworks
2gh search code "playwright OR puppeteer OR selenium path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 20

Integration with Dotfiles

Save Discovered Skills for Reference

Skills should be stored in your Dotfiles repo under ai/files/skills/. When discovering skills, save them to a research directory first:

bash
1# Use dotfiles environment variable (fallback to detection if not set)
2DOTFILES_ROOT="${DOTFILES_REPO_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || echo "${HOME}/Projects/dotfiles")}"
3SKILLS_ROOT="${DOTFILES_ROOT}/ai/files/skills"
4RESEARCH_DIR="${DOTFILES_ROOT}/ai/files/skills-research"
5
6# Create a skills research directory
7mkdir -p "${RESEARCH_DIR}"
8
9# Download and categorize interesting skills
10CATEGORY="terminal-management"
11mkdir -p "${RESEARCH_DIR}/${CATEGORY}"
12
13# Save search results
14gh search code "tmux OR zellij OR screen path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
15  --limit 30 \
16  --json repository,path,url > "${RESEARCH_DIR}/${CATEGORY}/search-results.json"
17
18# Extract and download top results
19jq -r '.[] | "\(.repository.fullName)|\(.path)|\(.url)"' \
20  "${RESEARCH_DIR}/${CATEGORY}/search-results.json" | \
21  head -5 | \
22  while IFS='|' read repo path url; do
23    echo "Downloading: ${repo} - ${path}"
24    gh api "repos/${repo}/contents/${path}" --jq '.content' | base64 -d > \
25      "${RESEARCH_DIR}/${CATEGORY}/$(echo ${repo} | tr '/' '-')-$(basename ${path})"
26  done

Suggest Storage Location

After downloading and reviewing a skill, suggest appropriate locations within ai/files/skills/:

bash
1# Function to suggest skill locations
2suggest_skill_location() {
3  local skill_name="$1"
4  local skill_content="$2"
5
6  echo "📂 Where should '${skill_name}' be stored in ai/files/skills/?"
7  echo ""
8
9  # Analyze content to suggest locations
10  if grep -qi "github\|gh cli\|pull request\|pr\|issue" "${skill_content}"; then
11    echo "1. ai/files/skills/github/${skill_name}/"
12    echo "   ✓ GitHub-related functionality"
13  fi
14
15  if grep -qi "test\|playwright\|selenium\|cypress" "${skill_content}"; then
16    echo "2. ai/files/skills/devtools/${skill_name}/"
17    echo "   ✓ Development and testing tools"
18  fi
19
20  if grep -qi "project\|task\|planning\|tracking" "${skill_content}"; then
21    echo "3. ai/files/skills/projectmanagement/${skill_name}/"
22    echo "   ✓ Project management and workflow"
23  fi
24
25  if grep -qi "design\|ui\|ux\|figma" "${skill_content}"; then
26    echo "4. ai/files/skills/design/${skill_name}/"
27    echo "   ✓ Design-related skills"
28  fi
29
30  if grep -qi "research\|search\|discover" "${skill_content}"; then
31    echo "5. ai/files/skills/research/${skill_name}/"
32    echo "   ✓ Research and discovery"
33  fi
34
35  echo ""
36  echo "Choose a number (1-5), or provide your own path:"
37  echo "Example: ai/files/skills/custom-category/${skill_name}/"
38}
39
40# Usage example
41SKILL_FILE="${RESEARCH_DIR}/terminal-management/example-skill.md"
42SKILL_NAME="zellij-manager"
43
44suggest_skill_location "${SKILL_NAME}" "${SKILL_FILE}"
45read -r location_choice
46
47case $location_choice in
48  1) TARGET_DIR="${SKILLS_ROOT}/github/${SKILL_NAME}" ;;
49  2) TARGET_DIR="${SKILLS_ROOT}/devtools/${SKILL_NAME}" ;;
50  3) TARGET_DIR="${SKILLS_ROOT}/projectmanagement/${SKILL_NAME}" ;;
51  4) TARGET_DIR="${SKILLS_ROOT}/design/${SKILL_NAME}" ;;
52  5) TARGET_DIR="${SKILLS_ROOT}/research/${SKILL_NAME}" ;;
53  *)
54    # User provided custom path
55    if [[ "$location_choice" =~ ^ai/files/skills/ ]]; then
56      TARGET_DIR="${DOTFILES_ROOT}/${location_choice}"
57    else
58      TARGET_DIR="${SKILLS_ROOT}/${location_choice}"
59    fi
60    ;;
61esac
62
63# Create target directory and move skill
64mkdir -p "${TARGET_DIR}"
65cp "${SKILL_FILE}" "${TARGET_DIR}/SKILL.md"
66echo "✅ Skill moved to: ${TARGET_DIR}"

Tips and Tricks

1. Fuzzy Matching with Wildcards

The regex in the path already handles variations like skill vs skills with the ? quantifier.

2. Combine with Other Search Terms

bash
1# Find skills that mention specific tools
2gh search code "gh cli OR 'github cli' path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 20

3. Language-Specific Searches

bash
1# Find skills written in specific languages with content search
2gh search code "language:markdown 'use when' path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
3  --limit 30

4. Exclude Certain Repos

bash
1# Search but exclude your own repos
2gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/ -user:yourusername" \
3  --limit 30

Error Handling

bash
1# Check if gh CLI is installed and authenticated
2if ! command -v gh &> /dev/null; then
3    echo "Error: GitHub CLI (gh) is not installed"
4    echo "Install with: brew install gh (macOS) or apt install gh (Linux)"
5    exit 1
6fi
7
8if ! gh auth status &> /dev/null; then
9    echo "Error: Not authenticated with GitHub"
10    echo "Run: gh auth login"
11    exit 1
12fi
13
14# Handle rate limiting
15gh search code "path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
16  --limit 30 || {
17    echo "Error: GitHub API rate limit exceeded"
18    echo "Check status: gh api rate_limit"
19    exit 1
20}

Best Practices

1. Start Broad, Then Narrow: Begin with general searches, then refine based on results
2. Check Multiple Frameworks: Skills from cursor might work in pi and vice versa
3. Look for Patterns: Similar problems often have similar solutions across repos
4. Respect Licenses: Check repository licenses before adapting code
5. Credit Sources: When adapting skills, note the original source
6. Update Your Own Skills: Share back improved versions to help others

Example: Complete Skill Discovery Session

bash
1#!/bin/bash
2# skill-discovery.sh - Discover and analyze skills for a specific topic
3
4TOPIC="${1:-testing}"
5
6# Use dotfiles environment variable
7DOTFILES_ROOT="${DOTFILES_REPO_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || echo "${HOME}/Projects/dotfiles")}"
8SKILLS_ROOT="${DOTFILES_ROOT}/ai/files/skills"
9RESEARCH_DIR="${DOTFILES_ROOT}/ai/files/skills-research"
10OUTPUT_DIR="${RESEARCH_DIR}/${TOPIC}"
11
12mkdir -p "${OUTPUT_DIR}"
13
14echo "🔍 Discovering skills related to: ${TOPIC}"
15echo "📍 Dotfiles: ${DOTFILES_ROOT}"
16echo "📂 Research output: ${OUTPUT_DIR}"
17echo ""
18
19# 1. Initial search
20echo "Step 1: Searching GitHub..."
21gh search code "${TOPIC} path:/^./.?(opencode|ai|llm|claude|codex|agents|pi|cursor|factory)\/skills?\/.*\.md$/" \
22  --limit 50 \
23  --json repository,path,url \
24  > "${OUTPUT_DIR}/search-results.json"
25
26RESULT_COUNT=$(jq length "${OUTPUT_DIR}/search-results.json")
27echo "Found ${RESULT_COUNT} results"
28
29# 2. Analyze repositories
30echo "Step 2: Analyzing repositories..."
31jq -r '.[].repository.fullName' "${OUTPUT_DIR}/search-results.json" | \
32  sort | uniq -c | sort -rn | \
33  head -10 > "${OUTPUT_DIR}/top-repos.txt"
34
35echo "Top repositories with ${TOPIC} skills:"
36cat "${OUTPUT_DIR}/top-repos.txt"
37
38# 3. Download top 3 skill files
39echo "Step 3: Downloading sample skills..."
40jq -r '.[0:3] | .[] | "\(.repository.fullName)|\(.path)"' \
41  "${OUTPUT_DIR}/search-results.json" | \
42  while IFS='|' read repo path; do
43    filename=$(echo "${repo}-${path}" | tr '/' '-')
44    echo "  Downloading: ${repo}/${path}"
45    gh api "repos/${repo}/contents/${path}" \
46      --jq '.content' | base64 -d \
47      > "${OUTPUT_DIR}/${filename}" 2>/dev/null || \
48      echo "    Failed to download"
49  done
50
51# 4. Create summary report
52echo "Step 4: Creating summary report..."
53cat > "${OUTPUT_DIR}/DISCOVERY-REPORT.md" << EOF
54# Skill Discovery Report: ${TOPIC}
55
56**Date**: $(date)
57**Total Results**: ${RESULT_COUNT}
58**Dotfiles Root**: ${DOTFILES_ROOT}
59
60## Top Repositories
61
62\`\`\`
63$(cat "${OUTPUT_DIR}/top-repos.txt")
64\`\`\`
65
66## All Results
67
68$(jq -r '.[] | "- [\(.repository.fullName)](\(.repository.url)) - [\(.path)](\(.url))"' \
69  "${OUTPUT_DIR}/search-results.json")
70
71## Storage Locations
72
73Review downloaded skills and determine appropriate location in:
74
75\`\`\`
76${SKILLS_ROOT}/
77├── github/          # GitHub-related functionality
78├── devtools/        # Development and testing tools
79├── projectmanagement/  # Project and task management
80├── design/          # Design and UI/UX tools
81├── research/        # Research and discovery
82├── superpowers/     # Advanced meta-skills
83├── experts/         # Domain expert personas
84└── [custom]/        # Your custom categories
85\`\`\`
86
87## Next Steps
88
891. Review downloaded skill files in: ${OUTPUT_DIR}
902. Identify common patterns and best practices
913. Choose appropriate location in ai/files/skills/
924. Adapt for your use case
935. Consider contributing improvements back
94
95## Move to Skills Directory
96
97\`\`\`bash
98# Example: Move adapted skill to appropriate location
99SKILL_NAME="my-new-skill"
100
101# Option 1: GitHub-related
102mkdir -p "${SKILLS_ROOT}/github/\${SKILL_NAME}"
103cp adapted-skill.md "${SKILLS_ROOT}/github/\${SKILL_NAME}/SKILL.md"
104
105# Option 2: DevTools
106mkdir -p "${SKILLS_ROOT}/devtools/\${SKILL_NAME}"
107cp adapted-skill.md "${SKILLS_ROOT}/devtools/\${SKILL_NAME}/SKILL.md"
108
109# Option 3: Custom location
110mkdir -p "${SKILLS_ROOT}/custom-category/\${SKILL_NAME}"
111cp adapted-skill.md "${SKILLS_ROOT}/custom-category/\${SKILL_NAME}/SKILL.md"
112\`\`\`
113
114EOF
115
116echo "✅ Discovery complete! Report saved to: ${OUTPUT_DIR}/DISCOVERY-REPORT.md"
117echo ""
118echo "Next: Review skills and move to ${SKILLS_ROOT}/"

Skill Storage Workflow

After discovering and adapting a skill, follow this workflow:

1. Analyze Skill Content

Review the skill to determine its primary purpose:

bash
1SKILL_FILE="downloaded-skill.md"
2
3echo "Analyzing skill content..."
4echo ""
5echo "Contains GitHub functionality:" $(grep -qi "github\|gh cli" "$SKILL_FILE" && echo "✓ Yes" || echo "✗ No")
6echo "Contains testing tools:" $(grep -qi "test\|playwright\|selenium" "$SKILL_FILE" && echo "✓ Yes" || echo "✗ No")
7echo "Contains project management:" $(grep -qi "project\|task\|planning" "$SKILL_FILE" && echo "✓ Yes" || echo "✗ No")
8echo "Contains design tools:" $(grep -qi "design\|ui\|figma" "$SKILL_FILE" && echo "✓ Yes" || echo "✗ No")
9echo "Contains research tools:" $(grep -qi "research\|search\|discover" "$SKILL_FILE" && echo "✓ Yes" || echo "✗ No")

2. Present Storage Options

Present 1-3 suggested locations based on content analysis:

📂 Where should this skill be stored in ai/files/skills/?

Suggested locations:

1. ai/files/skills/github/skill-name/
   ✓ Contains GitHub API usage
   ✓ Uses gh CLI extensively

2. ai/files/skills/research/skill-name/
   ✓ Primary purpose is discovery/research
   ✓ Searches external sources

3. ai/files/skills/devtools/skill-name/
   ✓ Developer tooling functionality
   ✓ Enhances development workflow

Choose a number (1-3), or provide your own path:
Example: ai/files/skills/custom-category/skill-name/

3. Interactive Selection Script

bash
1#!/bin/bash
2# store-skill.sh - Interactive skill storage helper
3
4SKILL_FILE="$1"
5SKILL_NAME="$2"
6
7if [[ -z "$SKILL_FILE" || -z "$SKILL_NAME" ]]; then
8  echo "Usage: $0 <skill-file> <skill-name>"
9  exit 1
10fi
11
12# Use dotfiles environment variable
13DOTFILES_ROOT="${DOTFILES_REPO_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || echo "${HOME}/Projects/dotfiles")}"
14SKILLS_ROOT="${DOTFILES_ROOT}/ai/files/skills"
15
16# Analyze content
17echo "🔍 Analyzing skill content..."
18echo ""
19
20SUGGESTIONS=()
21COUNT=1
22
23if grep -qi "github\|gh cli\|pull request\|issue" "$SKILL_FILE"; then
24  SUGGESTIONS+=("${COUNT}:github:GitHub-related functionality")
25  echo "${COUNT}. ai/files/skills/github/${SKILL_NAME}/"
26  echo "   ✓ GitHub-related functionality"
27  echo ""
28  ((COUNT++))
29fi
30
31if grep -qi "test\|playwright\|selenium\|cypress\|jest" "$SKILL_FILE"; then
32  SUGGESTIONS+=("${COUNT}:devtools:Testing and development tools")
33  echo "${COUNT}. ai/files/skills/devtools/${SKILL_NAME}/"
34  echo "   ✓ Testing and development tools"
35  echo ""
36  ((COUNT++))
37fi
38
39if grep -qi "project\|task\|planning\|tracking\|workflow" "$SKILL_FILE"; then
40  SUGGESTIONS+=("${COUNT}:projectmanagement:Project management")
41  echo "${COUNT}. ai/files/skills/projectmanagement/${SKILL_NAME}/"
42  echo "   ✓ Project management and workflow"
43  echo ""
44  ((COUNT++))
45fi
46
47if grep -qi "design\|ui\|ux\|figma\|sketch" "$SKILL_FILE"; then
48  SUGGESTIONS+=("${COUNT}:design:Design tools")
49  echo "${COUNT}. ai/files/skills/design/${SKILL_NAME}/"
50  echo "   ✓ Design and UI/UX tools"
51  echo ""
52  ((COUNT++))
53fi
54
55if grep -qi "research\|search\|discover\|analyze" "$SKILL_FILE"; then
56  SUGGESTIONS+=("${COUNT}:research:Research and discovery")
57  echo "${COUNT}. ai/files/skills/research/${SKILL_NAME}/"
58  echo "   ✓ Research and discovery"
59  echo ""
60  ((COUNT++))
61fi
62
63# If no suggestions, show common categories
64if [[ ${#SUGGESTIONS[@]} -eq 0 ]]; then
65  echo "No specific category detected. Common options:"
66  echo ""
67  echo "1. ai/files/skills/devtools/${SKILL_NAME}/"
68  echo "2. ai/files/skills/superpowers/${SKILL_NAME}/"
69  echo "3. ai/files/skills/experts/${SKILL_NAME}/"
70  echo ""
71fi
72
73echo "Choose a number, or provide your own path:"
74echo "Example: ai/files/skills/my-category/${SKILL_NAME}/"
75echo ""
76read -r -p "Selection: " choice
77
78# Process choice
79TARGET_DIR=""
80if [[ "$choice" =~ ^[0-9]+$ ]] && [[ "$choice" -le "${#SUGGESTIONS[@]}" ]]; then
81  # Numbered choice
82  CATEGORY=$(echo "${SUGGESTIONS[$((choice-1))]}" | cut -d: -f2)
83  TARGET_DIR="${SKILLS_ROOT}/${CATEGORY}/${SKILL_NAME}"
84elif [[ "$choice" =~ ^ai/files/skills/ ]]; then
85  # Full path provided
86  TARGET_DIR="${DOTFILES_ROOT}/${choice}"
87else
88  # Relative path assumed
89  TARGET_DIR="${SKILLS_ROOT}/${choice}"
90fi
91
92# Create and copy
93mkdir -p "${TARGET_DIR}"
94cp "${SKILL_FILE}" "${TARGET_DIR}/SKILL.md"
95
96echo ""
97echo "✅ Skill stored at: ${TARGET_DIR}/SKILL.md"
98echo ""
99echo "Next steps:"
100echo "1. Review and edit: ${TARGET_DIR}/SKILL.md"
101echo "2. Add frontmatter if needed (name, description)"
102echo "3. Commit to dotfiles repo"

4. Commit to Dotfiles

After storing the skill:

bash
1cd "${DOTFILES_ROOT}"
2git add "ai/files/skills/${CATEGORY}/${SKILL_NAME}/"
3git commit -m "feat(skills): add ${SKILL_NAME} skill
4
5Discovered from GitHub search, adapted for local use.
6Stored in ${CATEGORY} category."
7git push

See Also

• github: Core GitHub CLI skill
• writing-skills: Creating and editing skills
• superpowers/brainstorming: Design before implementation