mikeng-io/agent-skills/deep-explorer

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Git-based codebase exploration with delta analysis. Performs full exploration on first run, then incremental delta exploration tracking committed and uncommitted changes.

Install

npx vskill@latest install mikeng-io/agent-skills/deep-explorer

SKILL.md

Deep Explorer: Git-Based Codebase Exploration

Execute this skill to explore and understand how a codebase works, using Git to efficiently track changes between explorations.

Execution Instructions

When invoked, you will:

  1. Resolve scope and context — invoke context skill (always), then preflight if confidence is low
  2. Verify Git repository - Ensure codebase is a Git repository
  3. Determine exploration type - Full (first run) or Delta (incremental)
  4. Detect changes - Use Git to find what changed since last exploration
  5. Execute exploration - Explore entire codebase (full) or changed areas (delta)
  6. Generate report - Comprehensive exploration or delta change report
  7. Save baseline - Store current commit reference for next exploration

Dependency Check

Before executing any step, verify all required skills are present:

[skills-root]/context/SKILL.md
[skills-root]/preflight/SKILL.md
[skills-root]/domain-registry/README.md

Where [skills-root] is the parent of this skill's directory. Resolve with ls ../ from this skill's location.

If any required file is missing → stop immediately and output:

⚠ Missing required skills for deep-explorer:
{missing-skill}
Expected: {skills-root}/{missing-skill}/SKILL.md
Install the missing skill(s):
git clone https://github.com/mikeng-io/agent-skills /tmp/agent-skills
cp -r /tmp/agent-skills/skills/{missing-skill} {skills-root}/
Or install the full suite at once:
cp -r /tmp/agent-skills/skills/ {skills-root}/

All dependencies present → proceed to Step 0.


Step 0: Scope & Context Resolution

Context (always required):

Invoke Skill("context") first. It identifies what to explore, confirms artifact type, and establishes domain focus:

yaml
context_report:
artifact_type: "" # code | financial | marketing | creative | research | mixed
domains: [] # matched domain names from domain-registry
routing: "" # routing recommendation (informational for explorer)
confidence: "" # high | medium | low

Preflight (conditional — triggered by context confidence):

Invoke Skill("preflight") only if context_report.confidence == "low" OR one or more signals remain unresolved:

  • The repository or directory to explore is not clearly identified
  • The exploration intent is ambiguous (full map vs. specific area?)
  • Domains or focus areas cannot be inferred

Preflight fills exactly the gaps context could not resolve (max 3 questions, one at a time):

yaml
scope_clarification:
artifact: "" # repository or directory to explore
intent: "explore"
domains: [] # focus domains for exploration (supplements context_report.domains)
constraints: [] # specific areas to prioritize or exclude
confidence: "" # high | medium

If context_report.confidence == "high" → skip preflight entirely.

Merge into working scope:

yaml
working_scope:
artifact: "" # repository or directory to explore
domains: [] # from context_report, supplemented by preflight
constraints: [] # areas to prioritize or exclude
context_summary: "" # combined description for explorer agent prompts

Use working_scope to focus exploration agents on relevant domains and files.


Step 0: Verify Git Repository (Optional)

Check if the codebase is a Git repository to determine exploration mode:

bash
# Check if Git repository
if ! git rev-parse --is-inside-work-tree 2>/dev/null; then
# Git not available — proceed with full filesystem exploration
EXPLORATION_TYPE="full"
GIT_AVAILABLE=false
else
# Verify Git is functional (has at least one commit)
if ! git rev-parse HEAD 2>/dev/null; then
EXPLORATION_TYPE="full"
GIT_AVAILABLE=false
else
GIT_AVAILABLE=true
fi
fi

If not a git repository:

  • Set EXPLORATION_TYPE="full" (no delta mode without git)
  • Proceed with file system exploration instead
  • Note in report: "Git not available — full filesystem exploration mode"

Step 1: Determine Exploration Type

Check if previous exploration exists to determine strategy:

bash
# Check for previous exploration
LAST_EXPLORATION=".outputs/exploration/latest-exploration.json"
if [ -f "$LAST_EXPLORATION" ]; then
# Previous exploration exists
EXPLORATION_TYPE="delta"
BASELINE_COMMIT=$(jq -r '.baseline_commit' "$LAST_EXPLORATION")
BASELINE_TIMESTAMP=$(jq -r '.baseline_timestamp' "$LAST_EXPLORATION")
else
# No previous exploration
EXPLORATION_TYPE="full"
BASELINE_COMMIT=$(git rev-parse HEAD)
BASELINE_TIMESTAMP=$(date -u +"%Y-%m-%dT%H:%M:%SZ")
fi

Exploration Types:

FULL Exploration:

  • First run (no baseline)
  • Explore entire codebase
  • Establish baseline for future delta explorations

DELTA Exploration:

  • Subsequent runs (baseline exists)
  • Detect changes using Git
  • Re-explore only changed areas
  • Generate delta report with impact analysis

Step 2: Detect Changes (Delta Only)

For delta exploration, use Git to detect all changes:

Committed Changes
bash
# Get baseline commit from last exploration
BASELINE=$(jq -r '.baseline_commit' .outputs/exploration/latest-exploration.json)
CURRENT=$(git rev-parse HEAD)
# Find committed changes since baseline
git diff --name-status $BASELINE $CURRENT > committed_changes.txt
# Categorize changes
MODIFIED=$(git diff --name-only --diff-filter=M $BASELINE $CURRENT)
ADDED=$(git diff --name-only --diff-filter=A $BASELINE $CURRENT)
DELETED=$(git diff --name-only --diff-filter=D $BASELINE $CURRENT)
RENAMED=$(git diff --name-only --diff-filter=R $BASELINE $CURRENT)
Uncommitted Changes
bash
# Working directory changes (modified but not staged)
UNCOMMITTED_MODIFIED=$(git diff --name-only HEAD)
# Staged changes (committed to index but not committed)
STAGED=$(git diff --name-only --cached)
# Untracked files (new files not in Git)
UNTRACKED=$(git ls-files --others --exclude-standard)
Combine All Changes
yaml
changes_detected:
committed:
status: stable
files:
- modified: [list from git diff]
- added: [list from git diff]
- deleted: [list from git diff]
- renamed: [list from git diff]
uncommitted:
status: work_in_progress
files:
- modified: [list from git diff HEAD]
- staged: [list from git diff --cached]
- untracked: [list from git ls-files --others]
total_changes: <count>
exploration_scope: [list of unique files to re-explore]

Change Categorization:

  • Committed changes → Stable, unlikely to change
  • Uncommitted changes → Work in progress, may change
  • Untracked files → New additions, not yet in Git

Step 3: Execute Exploration

Full Exploration (First Run)

Spawn explorer sub-agents with dependency-aware execution for optimal efficiency:

Dependency-Aware Execution Strategy:

Exploration has natural dependencies - some tasks must complete before others:

  • Architecture analysis needs structure and technology information first
  • Workflow tracing needs structure and dependency information first
  • Technology and dependency analysis both need structure first

Task Definitions with Dependencies:

yaml
tasks:
# Wave 1: Foundation (must run first)
- id: structure
description: "Analyze directory structure"
agent: "structure-explorer"
depends_on: []
# Wave 2: Technology and dependency analysis (need structure)
- id: technology
description: "Identify technologies and frameworks"
agent: "technology-explorer"
depends_on: [structure]
- id: dependencies
description: "Map dependencies and imports"
agent: "dependency-explorer"
depends_on: [structure]
# Wave 3: Architecture and workflows (need structure + tech/deps)
- id: architecture
description: "Analyze architectural patterns"
agent: "architecture-explorer"
depends_on: [structure, technology]
- id: workflows
description: "Trace workflows and data flows"
agent: "workflow-explorer"
depends_on: [structure, dependencies]
execution:
mode: dag-orchestrated
waves:
wave_1: [structure] # 1 task
wave_2: [technology, dependencies] # 2 tasks in parallel
wave_3: [architecture, workflows] # 2 tasks in parallel
capability: high

Why This Order Matters:

  1. Structure first - You can't analyze architecture without knowing what files exist
  2. Tech and deps together - Both analyze different aspects of structure independently
  3. Architecture and workflows last - Both benefit from having structure + context

Performance Benefit:

  • Sequential execution: ~15 seconds (5 tasks × 3s avg)
  • Dependency-aware parallel: ~9 seconds (1 wave of 1 + 2 waves of 2)
  • Speedup: 1.7x

Explorer Agent Templates:

Structure Explorer Agent
Capability: high
You are a STRUCTURE EXPLORER. Analyze the repository structure and organization.
## Your Task
1. Map directory hierarchy using Glob and ls
2. Count files by type (extensions)
3. Identify organizational patterns
4. Find configuration files
5. Analyze naming conventions
## Output Format (JSON)
{
"agent": "structure-explorer",
"directory_structure": {
"total_directories": 24,
"total_files": 188,
"hierarchy": "..."
},
"file_distribution": {
"typescript": 145,
"javascript": 23,
"json": 12,
"markdown": 8
},
"organizational_patterns": [
"Feature-based organization",
"Separation of concerns",
"Test files colocated with source"
],
"configuration_files": [
"package.json",
"tsconfig.json",
".eslintrc.js",
"docker-compose.yml"
]
}
Technology Explorer Agent
Capability: high
You are a TECHNOLOGY EXPLORER. Inventory the technology stack.
## Your Task
1. Detect languages from file extensions
2. Identify frameworks from package.json, requirements.txt, etc.
3. Find build tools (Makefile, scripts, etc.)
4. Extract dependencies and versions
## Output Format (JSON)
{
"agent": "technology-explorer",
"languages": [
{"name": "TypeScript", "file_count": 145, "primary": true},
{"name": "JavaScript", "file_count": 23, "primary": false}
],
"frameworks": [
{"name": "Express.js", "version": "^4.18.0", "purpose": "web server"},
{"name": "React", "version": "^18.2.0", "purpose": "frontend"}
],
"build_tools": ["npm", "webpack", "docker"],
"dependencies": {
"production": {...},
"development": {...}
}
}
Architecture Explorer Agent
Capability: high
You are an ARCHITECTURE EXPLORER. Identify architectural patterns and structure.
## Your Task
1. Analyze project structure (MVC, microservices, etc.)
2. Identify module organization
3. Detect design patterns used
4. Assess separation of concerns
## Output Format (JSON)
{
"agent": "architecture-explorer",
"project_structure": "Layered architecture",
"layers": [
{"name": "API Layer", "location": "src/api/"},
{"name": "Business Logic", "location": "src/services/"},
{"name": "Data Layer", "location": "src/models/"}
],
"design_patterns": [
{"pattern": "Factory", "location": "src/factories/"},
{"pattern": "Repository", "location": "src/repositories/"},
{"pattern": "Dependency Injection", "implementation": "constructor-based"}
],
"separation_quality": "high"
}
Workflow Explorer Agent
Capability: high
You are a WORKFLOW EXPLORER. Trace code flows and entry points.
## Your Task
1. Find entry points (main files)
2. Trace key workflows
3. Map data flow patterns
4. Identify integration points

This is the start of the file. Read the full SKILL.md on GitHub.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
6
7-day trend score
10

Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 27 Mar 2026, 7 months ago.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/mikeng-io/agent-skills/deep-explorer/badge)](https://verified-skill.com/skills/mikeng-io/agent-skills/deep-explorer)

Something wrong with this skill or its page? Tell us.