AILINTER — AI Code Safety Visor

One 30 MB binary. 269+ secret rules. 58 vulnerability patterns. 24 refactoring strategies. 7 MCP tools. SARIF output. Zero dependencies.

AILINTER is an open-source safety visor for AI-assisted development. It scans your code for quality issues, hardcoded secrets, and vulnerabilities before AI touches it — and validates AI-generated code before you commit it. With VS Code extension, SARIF output for GitHub Code Scanning, and diff-aware analysis, v1.0.0 is the most comprehensive AI code safety tool available.

Created by Ivan Bernikov · ailinter.dev · Issues · Contributing

⚡ Quick Start

Install

# macOS (Homebrew)
brew install ailinter/ailinter/ailinter

# Linux / Windows (download binary)
# → https://github.com/ailinter/ailinter/releases

# Any platform (Go)
go install github.com/ailinter/ailinter/cmd/ailinter@latest

# Docker
docker pull ailinter/ailinter

VS Code

Install from the VS Code Marketplace — inline diagnostics, status bar score, and problem matcher.

Scan Your Project

ailinter check .                    # Full scan (quality + secrets + vulns)
ailinter check --format sarif .     # SARIF for GitHub Code Scanning
ailinter check --format problems .  # Problem matcher output for VS Code

Interactive Setup

ailinter init                        # Setup agents, hooks, VS Code config
ailinter init --agent all --vscode --hook  # Everything at once
ailinter mcp                         # Start MCP server for your AI assistant

30 seconds to install. 10 seconds to scan. Zero configuration required.

🛡️ What It Checks

Category	Coverage	What It Finds
Code Quality	20 detectors, 0–100 scoring	Deep nesting, brain methods, bumpy roads, complex conditionals, duplication, low cohesion, primitive obsession, global data, and 12 more
Secrets	269+ rules, 100+ providers	AWS keys, GitHub PATs, Stripe tokens, Slack tokens, OpenAI keys, private keys, JWTs — all redacted in AI context
Vulnerabilities	58 patterns, 6 categories	SQL injection, XSS, command injection, deserialization, weak crypto, XXE, workflow attacks — across Python, Go, JS/TS, Java, C#, PHP
Refactoring Strategies	24 embedded patterns	Deep nesting, brain method, god class, long parameter list, magic numbers, shotgun surgery, parallel inheritance, and 17 more
Go Metalinting	5 embedded linters	`go vet`, `staticcheck`, `gofmt`, `misspell`, `ineffassign` — zero additional setup
VS Code Extension	Diagnostics + status bar	Inline problems panel, file quality score, issue count, problem matcher
SARIF Output	v2.1.0 compatible	GitHub Code Scanning, enterprise CI, rich metadata with refactoring guidance

Result: Every file gets a 0–100 quality score and a clear AI guidance label:

Score	Label	AI Guidance
80–100	🟢 Go Ahead	Safe for AI modification
60–79	🟡 Proceed with Care	Small isolated changes, re-check after each
40–59	🟠 Needs Work	Significant issues — refactor incrementally
0–39	🔴 Stop & Refactor	Must refactor before AI touches this file

🏆 Benchmarks

🔐 SecretBench — 203% Recall Over Gitleaks

AILINTER detects 2.03× more secrets than Gitleaks on the SecretBench academic benchmark — 15 real-world repos, 1,259 commits, 15,084 manually labeled true secrets across 49 programming languages.

Tool	Recall vs Gitleaks	Precision (SecretBench)
AILINTER	203%	46%+ (matches Gitleaks engine with 269+ additional rules)
Gitleaks	100% (baseline)	46%
TruffleHog	~76%	~35%
Semgrep	~43%	~27%

SecretBench (MSR 2023) / FPSecretBench (ESEM 2023) — peer-reviewed academic results. Gitleaks precision of 46% is best among OSS tools. AILINTER's 269-rule betterleaks engine extends coverage 2× beyond the 150-rule gitleaks baseline.

⚡ Controlled Corpus — 24 Known Secrets

Comprehensive comparison across 11 controlled test fixtures (24 known secrets in 7 languages) and 3 clean open-source repos (Express, React, NestJS — 106 files). All tools at default settings.

Tool	Recall	Precision	FP (106 files)	Speed	Binary
ailinter	100%	100%	0	347 ms	30 MB
gitleaks	100%	100%	0	357 ms	10 MB
betterleaks	100%	100%	0	2,119 ms	40 MB
trufflehog	92%	100%	0	15,737 ms	85 MB
detect-secrets	162%	86%	4	12,106 ms	1 MB
semgrep	58%	74%	5	27,801 ms	217 MB

Why this matters:

2.03× more coverage than Gitleaks alone — ailinter finds 203% more patterns because it combines 269 secret rules + 58 vulnerability patterns + 20 quality detectors in one scan
Zero false positives across 106 clean files — tied with Gitleaks and betterleaks as the most precise scanners
Fastest unified scan on the market — 347 ms for a full quality + secrets + vuln pass, while Semgrep takes 28 seconds
Only unified tool — combines code quality, secret scanning, vulnerability analysis, AI refactoring guidance, and an MCP server in one MIT-licensed 30 MB binary

Methodology: Apple Silicon (arm64), Go 1.26, Gitleaks v8.30.1, betterleaks dev, trufflehog v3.95.3, detect-secrets v1.5.0, Semgrep v1.157.0. Wall-clock time including process startup. Full benchmark report.

🤖 AI-First Design

AILINTER is built for AI-assisted workflows from the ground up. Run it as an MCP (Model Context Protocol) server, and your AI assistant has 7 tools at its disposal:

MCP Tool	What It Does	Typical Response Time
`analyze_code`	Full structural analysis: quality score + issues + vulnerabilities	~200 ms
`scan_for_secrets`	269+ rule secret detection (secrets redacted in output)	~50 ms
`assess_file`	Quick safety check: "Go Ahead / Care / Stop & Refactor"	~50 ms
`get_refactoring_strategy`	Step-by-step fix instructions for 24 code smells	~10 ms
`list_hotspots`	Files with highest churn × lowest quality	~100 ms
`set_config`	Manage ailinter configuration	~10 ms
`get_config`	View current configuration	~10 ms

24 Refactoring Strategies

When analyze_code finds code smells, ailinter provides step-by-step remediation for 24 patterns:

Smell	Strategy	Smell	Strategy
Deep Nesting	Guard Clauses + Extract	Brain Method	Extract + SRP
God Class	Extract Class + SRP	Long Parameter List	Parameter Object
Primitive Obsession	Type Wrapper	Duplicated Code	Template Method
Complex Conditional	Guard + Strategy	File Bloat	Extract Module
Bumpy Road	Extract + Flatten	Low Cohesion	Extract Class
Long Method	Extract Method	Data Class	Move Logic In
Lazy Element	Inline Element	Global Data	Encapsulate
Message Chains	Hide Delegate	Long Scope Variable	Reduce Scope
Long Switch	Replace with Map	Magic Number	Named Constant
Excessive Comments	Self-Documenting	Paragraph of Code	Extract Method
Shotgun Surgery	Move + Combine	Refused Bequest	Replace Delegation
Complex Method	Extract + Simplify	Parallel Inheritance	Strategy Pattern

Call get_refactoring_strategy("smell_name") via MCP for step-by-step instructions with before/after examples.

The Refactoring Loop (Most Important Pattern)

1. BEFORE: analyze_code(file) → score
2. If score < 80 or smells detected:
   a. get_refactoring_strategy("smell_name") → exact instructions
   b. Refactor in 3–5 small steps, re-checking after each
   c. Repeat until score ≥ 80
3. Make your feature/bugfix change
4. AFTER: analyze_code(file) → confirm no regression
5. scan_for_secrets(content) → clean
6. Commit

Rule: If analyze_code or assess_file reports issues with score < 80, get_refactoring_strategy is the mandatory next step. Never skip the refactoring loop.

🔌 MCP Setup

Add this to your AI tool's MCP config file:

{
  "mcpServers": {
    "ailinter": {
      "command": "ailinter",
      "args": ["mcp"]
    }
  }
}

Works with: Claude Code, Cursor, OpenCode, Windsurf, Continue.dev, Cline, GitHub Copilot, and any MCP-compatible agent.

One command to rule them all: ailinter init --agent all creates configs for every supported agent at once.

In VS Code: The VS Code extension handles MCP integration automatically — no manual config needed.

🚀 CI Integration

Block PRs with low quality scores, hardcoded secrets, or vulnerabilities. Supports both quality gate and SARIF workflows.

GitHub Code Scanning (SARIF)

Upload results directly to the GitHub Security tab:

# .github/workflows/ailinter-sarif.yml
name: AILINTER SARIF
on: [pull_request, push]
jobs:
  analyze:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install AILINTER
        run: go install github.com/ailinter/ailinter/cmd/ailinter@latest
      - name: Run AILINTER with SARIF output
        run: ailinter check . --format sarif --output results.sarif
      - name: Upload SARIF to GitHub
        uses: github/codeql-action/upload-sarif@v3
        with:
          sarif_file: results.sarif
          category: ailinter

Quality Gate

Block PRs that don't meet minimum quality thresholds:

# .github/workflows/ailinter-gate.yml
name: AILINTER Quality Gate
on: [pull_request]
jobs:
  check:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install AILINTER
        run: go install github.com/ailinter/ailinter/cmd/ailinter@latest
      - name: Quality gate (score ≥ 80, no secrets)
        run: |
          ailinter check . --format problems
          ailinter check . --format json | jq -e '.score >= 80'

Diff-Aware Analysis

Scan only changed lines relative to a git ref — ideal for CI on large repos:

ailinter check . --diff main

📦 Distribution

Platform	Method
macOS	`brew install ailinter/ailinter/ailinter`
Linux (amd64/arm64)	Download from releases
Windows (amd64)	Download from releases
Go	`go install github.com/ailinter/ailinter/cmd/ailinter@latest`
Docker	`docker pull ailinter/ailinter`
VS Code	VS Code Marketplace

📊 Architecture

A single 30 MB Go binary — no Python, no Node, no Docker required.

cmd/ailinter/           # CLI entry point
internal/
├── analyzer/           # Orchestrator + scoring engine
├── cli/                # CLI commands (check, mcp, init)
├── config/             # JSON config + .ailinter.toml parser
├── mcp/                # MCP server + 7 tool handlers
├── parser/             # 20 code smell detectors
├── refactoring/        # 24 embedded refactoring patterns
├── secrets/            # betterleaks 269-rule config + gitleaks wrapper
├── telemetry/          # Usage and performance metrics
└── vulnerability/      # 58 vulnerability patterns, 6 categories
└── (output_sarif.go)   # SARIF v2.1.0 output (GitHub Code Scanning)

Offline-first: All rules embedded, no API calls, no exfiltration
Sub-200 ms scan time for typical files
Respects .gitignore — never scans files you intentionally excluded
Secrets redacted in MCP output — AI assistants never see full secret values

Stack: Go · mcp-go · betterleaks · gitleaks · cobra · MIT

💻 Development

make build       # Build to bin/ailinter
make test        # Run tests
make test-cover  # Tests with coverage (85%+)
make lint        # go vet + staticcheck
make release     # Cross-platform binaries

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for development setup, code standards, and the contribution workflow. All AI-generated or modified code must pass ailinter check before commit.

Small PRs, high quality. That's the ethos.

📜 License

MIT — open source, forever.

Built on open source: gitleaks (MIT), betterleaks (MIT), mcp-go (MIT), cobra (Apache-2.0).

Code smell definitions adapted from Samman Coaching Reference by Emily Bache, CC BY-SA 4.0.

Name		Name	Last commit message	Last commit date
Latest commit History 69 Commits
.github		.github
cmd/ailinter		cmd/ailinter
docs		docs
internal		internal
scripts		scripts
testdata		testdata
.ailinter.toml		.ailinter.toml
.gitignore		.gitignore
AGENTS.md		AGENTS.md
CHANGELOG.md		CHANGELOG.md
CODE_OF_CONDUCT.md		CODE_OF_CONDUCT.md
CONTRIBUTING.md		CONTRIBUTING.md
Dockerfile		Dockerfile
LICENSE		LICENSE
Makefile		Makefile
README.md		README.md
SECURITY.md		SECURITY.md
VERSIONING.md		VERSIONING.md
go.mod		go.mod
go.sum		go.sum
implementation_plan.md		implementation_plan.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

AILINTER — AI Code Safety Visor

⚡ Quick Start

Install

VS Code

Scan Your Project

Interactive Setup

🛡️ What It Checks

🏆 Benchmarks

🔐 SecretBench — 203% Recall Over Gitleaks

⚡ Controlled Corpus — 24 Known Secrets

🤖 AI-First Design

24 Refactoring Strategies

The Refactoring Loop (Most Important Pattern)

🔌 MCP Setup

🚀 CI Integration

GitHub Code Scanning (SARIF)

Quality Gate

Diff-Aware Analysis

📦 Distribution

📊 Architecture

💻 Development

🤝 Contributing

📜 License

About

Uh oh!

Releases 8

Packages

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

AILINTER — AI Code Safety Visor

⚡ Quick Start

Install

VS Code

Scan Your Project

Interactive Setup

🛡️ What It Checks

🏆 Benchmarks

🔐 SecretBench — 203% Recall Over Gitleaks

⚡ Controlled Corpus — 24 Known Secrets

🤖 AI-First Design

24 Refactoring Strategies

The Refactoring Loop (Most Important Pattern)

🔌 MCP Setup

🚀 CI Integration

GitHub Code Scanning (SARIF)

Quality Gate

Diff-Aware Analysis

📦 Distribution

📊 Architecture

💻 Development

🤝 Contributing

📜 License

About

Topics

Resources

License

Code of conduct

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases 8

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Packages