Cpg Analysis logo

Cpg Analysis

Community
alinaqi
cpg-analysis

Deep code property graph analysis with Joern CPG (AST+CFG+PDG) and CodeQL for control flow, data flow, taint analysis, and security auditing

Overview

Publisheralinaqi
Repositorymaggy
Skill namecpg-analysis
Stars
707
Forks
57
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by alinaqi on GitHub. Read the source before you install it.

Installation

Install the Cpg Analysis AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/alinaqi/maggy/tree/main/skills/cpg-analysis
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/alinaqi/maggy.git /tmp/maggy
mkdir -p .claude/skills
cp -r /tmp/maggy/skills/cpg-analysis .claude/skills/cpg-analysis
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cpg Analysis in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cpg Analysis on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cpg Analysis is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

CPG Analysis Skill

Purpose: Deep code analysis beyond AST. Use Joern for full Code Property Graph (control flow, data flow, program dependencies) and CodeQL for interprocedural taint analysis and vulnerability detection.

These are opt-in tools. They require Docker/JVM (Joern) or CodeQL CLI. Use codebase-memory-mcp (Tier 1, always-on) for everyday navigation. Use these for deep analysis when Tier 1 is not enough.

┌────────────────────────────────────────────────────────────────┐
│  CODE PROPERTY GRAPH = AST + CFG + CDG + DDG + PDG             │
│  ─────────────────────────────────────────────────────────────│
│  AST  = Abstract Syntax Tree (structure)                       │
│  CFG  = Control Flow Graph (execution paths)                   │
│  CDG  = Control Dependency Graph (conditional dependencies)    │
│  DDG  = Data Dependency Graph (data flow between statements)   │
│  PDG  = Program Dependency Graph (CDG + DDG combined)          │
│                                                                │
│  Tier 2 (Joern): Full CPG with 40+ query tools                │
│  Tier 3 (CodeQL): Interprocedural taint + security queries     │
└────────────────────────────────────────────────────────────────┘

Tier Selection Guide

Simple symbol lookup, dependency trace, blast radius?
  → Tier 1: codebase-memory-mcp (always on, sub-ms)

Control flow paths, data flow, dead code, complex refactoring?
  → Tier 2: Joern CPG (on-demand, seconds)

Security audit, taint analysis, vulnerability detection?
  → Tier 3: CodeQL (on-demand, seconds to minutes)

Full security review before release?
  → All three tiers in sequence

Tier 2: Joern CPG (CodeBadger MCP)

When to Use Joern

ScenarioWhy JoernTier 1 Can't Do This
Trace data flow through functionsFull DDG traversalTier 1 has no data flow
Understanding control flow pathsCFG analysis with branch conditionsTier 1 has no CFG
Finding dead/unreachable codePDG reachability analysisTier 1 only detects unused exports
Complex refactoring impactCross-function dependency chainsTier 1 limited to call graph
Auditing third-party library usageDeep call chain traversalTier 1 stops at import boundary
Understanding exception flowCFG includes throw/catch pathsTier 1 ignores exceptions

Key MCP Tools (Joern/CodeBadger)

ToolPurposeExample Query
generate_cpgBuild CPG for projectFirst-time setup or after major changes
get_cpg_statusCheck CPG build statusVerify CPG is ready before querying
run_cpgql_queryRun arbitrary CPGQL queriescpg.method("login").callOut.code.l
get_cpgql_syntax_helpQuery language referenceWhen unsure about query syntax
get_cfgControl flow graph for a methodUnderstand execution paths in a function
list_methodsList all methods in projectOverview of available functions
get_method_sourceGet source code of a methodRead specific function source
list_callsList calls from/to a methodCaller/callee analysis
get_call_graphFull call graph visualizationUnderstand call chains
get_type_definitionType/class definitionsUnderstand type hierarchy

Supported Languages (Joern)

Java, Scala, C/C++, Python, JavaScript, TypeScript, PHP, Ruby, Go, Kotlin, Swift, Lua

Not supported: Rust (use CodeQL for Rust)

MCP Configuration (Joern)

json
{
  "mcpServers": {
    "codebadger": {
      "url": "http://localhost:4242/mcp",
      "type": "http"
    }
  }
}

Prerequisites

  • Docker (for Joern backend)
  • Python 3.10+ (for MCP server)
  • Install: ~/.claude/install-graph-tools.sh --joern

Common CPGQL Queries

scala
// Find all methods that handle user input
cpg.method.where(_.parameter.name(".*input.*|.*request.*")).name.l

// Trace data flow from parameter to return
cpg.method("processPayment").parameter.reachableBy(cpg.method("processPayment").methodReturn).l

// Find methods with high cyclomatic complexity
cpg.method.where(_.controlStructure.size > 10).name.l

// Dead code: methods with no callers
cpg.method.where(_.callIn.size == 0).filter(_.name != "main").name.l

// Exception flow: methods that can throw but callers don't catch
cpg.method.where(_.ast.isThrow.size > 0).callIn.method.filter(_.ast.isTry.size == 0).name.l

Tier 3: CodeQL

When to Use CodeQL

ScenarioWhy CodeQLOther Tiers Can't Do This
Security audit before releaseInterprocedural taint analysisJoern has basic taint, CodeQL is deeper
Reviewing auth/payment codeData flow from source to sinkCross-function, cross-file taint
PR security reviewTargeted vulnerability scanPre-built OWASP query packs
Compliance checkingCWE/OWASP pattern matchingCurated security query suites
Rust security analysisFull Rust supportJoern doesn't support Rust

Key MCP Tools (CodeQL)

ToolPurpose
run_queryExecute a CodeQL query against the database
find_definitionsLocate symbol definitions
find_referencesFind all references to a symbol
get_resultsParse BQRS (Binary Query Result Sets)

Supported Languages (CodeQL)

C/C++, C#, Go, Java, Kotlin, JavaScript, TypeScript, Python, Ruby, Swift, Rust

MCP Configuration (CodeQL)

json
{
  "mcpServers": {
    "codeql": {
      "command": "codeql-mcp",
      "args": ["--database", ".code-graph/codeql-db"]
    }
  }
}

Prerequisites

  • CodeQL CLI (brew install codeql on macOS)
  • Install: ~/.claude/install-graph-tools.sh --codeql

Common CodeQL Patterns

ql
// SQL injection: user input flows to SQL query
import python
from DataFlow::PathNode source, DataFlow::PathNode sink
where TaintTracking::hasFlowPath(source, sink)
  and source instanceof RemoteFlowSource
  and sink instanceof SqlExecution
select sink, source, sink, "SQL injection from $@.", source, "user input"

// Unvalidated redirect
from DataFlow::PathNode source, DataFlow::PathNode sink
where source instanceof RemoteFlowSource
  and sink instanceof RedirectSink
select sink, "Unvalidated redirect from user input"

Combined Workflow: Deep Analysis

When performing security review or complex refactoring, use all tiers:

1. SCOPE       → Tier 1: detect_changes / get_architecture
                 Identify files and modules in scope

2. STRUCTURE   → Tier 1: search_graph / trace_call_path
                 Map the call graph and dependencies

3. FLOW        → Tier 2: get_cfg / run_cpgql_query
                 Analyze control flow and data flow paths

4. SECURITY    → Tier 3: run_query with taint analysis
                 Check for vulnerabilities in data paths

5. REPORT      → Combine findings from all tiers
                 Prioritize: Critical > High > Medium > Low

Anti-Patterns

Anti-PatternDo This Instead
Using Joern/CodeQL for simple symbol lookupUse Tier 1 search_graph (sub-ms vs seconds)
Running full CPG build on every commitBuild CPG on-demand; use Tier 1 for continuous monitoring
Querying Joern without checking get_cpg_statusAlways verify CPG is built and current before querying
Running CodeQL without a specific security questionHave a hypothesis first; CodeQL queries are expensive
Ignoring Tier 1 blast radius before deep analysisAlways scope with Tier 1 first, then go deep on flagged areas
Using CodeQL for non-security structural queriesUse Joern CPGQL for structural/flow queries; CodeQL for security

Frequently asked questions

What does the Cpg Analysis AI skill do?

Deep code property graph analysis with Joern CPG (AST+CFG+PDG) and CodeQL for control flow, data flow, taint analysis, and security auditing

Why use Cpg Analysis on TypingMind?

Because you install it once and use it with any model. Cpg Analysis is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cpg Analysis in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/alinaqi/maggy/tree/main/skills/cpg-analysis. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Cpg Analysis?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cpg Analysis?

As many as you like. As long as a model supports skills, you can use Cpg Analysis with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cpg Analysis AI skill free?

Yes. It is published on GitHub by alinaqi under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇