nika

[!TIP] Fun fact: The name Nika actually comes from the anime One Piece.

Nika is an open-source source code review and static analysis tool for security engineers who need to identify exploit paths in Java microservices. It performs cross-file taint analysis to trace attacker-controlled input across application layers and determine whether that input reaches a security-sensitive sink.

Key Capabilities

• Cross-file taint analysis
• AI exploitability analysis
• Branch-aware scanning
• Customizable rules

Why Nika

Many exploitable issues are not visible inside a single file. Request data may enter through a controller, pass through DTOs and service layers, and only become dangerous when it reaches a sink such as a database query, file operation, template engine, reflection API, or outbound network call.

Nika is built for that review problem. Instead of just identifying dangerous sinks, it traces data flow across files and functions so security engineers can determine whether a path is actually reachable.

What Nika Helps Security Engineers Do

• Trace attacker-controlled input across controllers, services, helpers, and utility layers.
• Validate source-to-sink reachability.
• Support secure code review with branch-aware scanning.
• Generate HTML reports.
• Extend coverage with custom sources, OpenGrep sinks, and vulnerability plugins.

Detection Coverage

Nika currently supports the following vulnerability categories:

• SQL injection
• SSRF
• Path traversal
• Command injection
• Code injection
• Template injection
• Deserialization
• XXE
• Cryptographic failures
• Unsafe reflection
• Security-critical call-order violations in sensitive execution flows and validation chains

How Nika Works

At a high level, Nika follows this analysis flow:

1. Process the target repository into an analysis representation that captures code structure, control flow, and data flow.
2. Identify configured sources where attacker-controlled input enters the application.
3. Identify sinks that represent security-sensitive operations.
4. Perform cross-file and inter-procedural analysis to determine whether input can reach those sinks.
5. Optionally review vulnerability with AI Agent to reduce false positives.
6. Produce an HTML report with the vulnerable path, affected code locations, and remediation context.

Quick Start

Run via Docker

You can use pre-built docker images.

git clone https://github.com/PhonePe/nika.git
docker pull ghcr.io/phonepe/nika:latest
export NIKA_IMAGE=ghcr.io/phonepe/nika
./run.sh --path /absolute/path/to/code --config /absolute/path/to/crtConfig.yml --output ./report.html

You still need to clone the repository after pulling the image, since run.sh makes it easier to run Nika by handling multiple mounts for you.

or build a docker image yourself.

git clone https://github.com/PhonePe/nika.git
cd nika
./build.sh
./run.sh --path /absolute/path/to/code --config /absolute/path/to/crtConfig.yml --output ./report.html

Run locally

git clone https://github.com/PhonePe/nika.git
cd nika
./native-build.sh
./native-run.sh --path /absolute/path/to/code --output ./report.html

Enable AI based False Positive Analysis

• Docker Setup - You can modify the config at /absolute/path/to/crtConfig.yml.
• Local Setup - You can modify the config at /absolute/path/to/native-crtConfig.yml.

LLMConfig:
  API_KEY: 'API_TOKEN'
  LLM_URL: 'https://chatgpt.com/api/v1'
  MODEL: 'GPT-5'
  MAX_TOOL_CALLS: 10
  MAX_ITERATIONS: 15
  RECURSION_LIMIT: 100
  PROMPT_COST_PER_MILLION: 1.25
  COMPLETION_COST_PER_MILLION: 10.0

llm_review_enabled: false

Documentation

To know more about Nika and it's features, you can read our detailed documentation here.

Language Support

Java is the only fully supported language today; support for other languages remains planned.

Contributors