Security tooling has always been fragmented. SAST scanners, SCA analyzers, container auditors, secret detectors, IaC linters, DAST engines, cloud posture assessments — each one lives in its own silo with its own CLI, configuration format, and output schema. What if your AI coding assistant could run all of them on demand, interpret the results, and deliver actionable remediation guidance?
That’s exactly what Product Security AI Skills delivers. It’s an open-source repository of 36 ready-to-use AI agent skills spanning the complete application security toolchain — built for Claude Code, Cursor, GitHub Copilot, and any AI coding assistant that supports the skills format.
Quick Start
Install all 36 skills in one command:
Or install individual skills as needed:
All skills are also available on the Vercel Skills Store — browse, search, and install directly from the web.
16 Security Categories, 36 Skills
Every major security domain is covered — from code-level analysis to cloud infrastructure posture. Each skill wraps a best-in-class open-source tool, handles execution, parses output, and delivers findings with remediation guidance.
Secure Coding
2 skillsTools: OWASP Rules | Targets: All (Rule-based & Gen)
SAST
10 skillsTools: Semgrep, Bandit, ESLint, SpotBugs, gosec, Flawfinder, Brakeman, Psalm, cargo-audit, detekt | Targets: 30+ Languages: Python, JS/TS, Java, Go, C/C++, Ruby, PHP, Rust, Kotlin
SCA
4 skillsTools: OSV-Scanner, Grype, npm audit, pip-audit | Targets: All Ecosystems: Node.js, Python, npm, PyPI, Maven
Secret Scanning
2 skillsTools: Gitleaks, TruffleHog | Targets: Git Repos, Files, S3
Container Security
3 skillsTools: Trivy, Hadolint, Dockle | Targets: Docker/OCI Images, Dockerfiles, CIS Benchmarks
Infrastructure as Code
3 skillsTools: Checkov, tfsec, KubeLinter | Targets: Terraform, K8s, Helm, CloudFormation
DAST
2 skillsTools: OWASP ZAP, Nuclei | Targets: Web Apps, APIs, Network
API Security
2 skillsTools: Schemathesis, Spectral | Targets: OpenAPI, GraphQL, AsyncAPI
SBOM
1 skillTools: Syft | Targets: Images, Filesystems
License Compliance
1 skillTools: ScanCode Toolkit | Targets: Source Code
Cloud Security
2 skillsTools: Prowler, ScoutSuite | Targets: AWS, Azure, GCP, Oracle
Mobile Security
1 skillTools: MobSF | Targets: Android, iOS
Network Security
1 skillTools: Nmap | Targets: Hosts, Networks
TLS/SSL Security
1 skillTools: testssl.sh | Targets: TLS Endpoints
Malware Analysis
1 skillTools: YARA | Targets: Files, Binaries
Supply Chain Security
1 skillTools: Confused + GuardDog | Targets: npm, PyPI, Maven
How It Works
Each skill is a self-contained SKILL.md file that teaches your AI agent how to:
- >Execute the underlying security tool with the right flags and configuration
- >Parse and interpret tool output (JSON, SARIF, or custom formats)
- >Classify findings by severity and provide contextual remediation guidance
- >Handle edge cases — missing dependencies, unsupported file types, empty results
Skills are available in the skills.sh format (top-level directories with SKILL.md files) and as native Claude Code skills in the .claude/skills/ directory.
OWASP Secure Coding Integration
The secure-coding-audit and secure-coding-generate skills are powered by the companion OWASP Secure Coding Practices (Markdown) repository — 22 modular rule files covering authentication, input validation, API security, cryptography, Docker, Kubernetes, CI/CD, and more.
$ cp -r owasp-secure-coding-md/rules ./rules
All other skills function independently without external dependencies.
Why This Matters
The shift to AI-assisted development is accelerating. Developers are using Claude Code, Cursor, and Copilot to write code faster than ever — but security tooling hasn’t kept pace. Most security scanning still happens in CI/CD pipelines, after code is committed.
Product Security AI Skills brings security scanning into the development loop. Instead of waiting for pipeline results, developers can run SAST, SCA, secret scanning, and container audits directly from their AI assistant — catching vulnerabilities at the moment of creation.
- >Shift-left security without changing developer workflows
- >Consistent security coverage across 30+ programming languages
- >AI-interpreted results with actionable remediation, not raw tool dumps
- >Open-source tools only — no vendor lock-in, no license fees
Get Product Security AI Skills
36 security skills. 16 categories. One install command. Secure your software supply chain with AI-powered precision.
View on GitHubOpen-source and free to use. Stars and contributions welcome.
