agent-security-scanning

Scan AI-generated code and dependencies for OWASP LLM vulnerabilities and CVEs.

31|8|Updated Apr 10, 2026

Search Tags:#static analysis #owasp #security scanning #dependency audit #prompt injection #cve #agent threat modeling

Ready to run

npx skills add https://github.com/itallstartedwithaidea/agent-skills --skill agent-security-scanning

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-security-scanning
Source: https://github.com/itallstartedwithaidea/agent-skills/tree/main/skills/security/agent-security-scanning
Command: npx skills add https://github.com/itallstartedwithaidea/agent-skills --skill agent-security-scanning

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent Security Scanning reduces the risk of shipping exploitable vulnerabilities hidden in AI-generated code by running targeted security analysis before production.

Core Features & Use Cases

OWASP Top 10 for LLM Applications coverage: Flags common LLM/agent-specific weaknesses and insecure patterns (e.g., prompt injection vectors, tool misuse, and data exfiltration paths).
Static code scanning (injection-focused): Identifies vulnerability patterns tied to agent-generated behavior such as unsanitized dynamic SQL, XSS/SSRF/command injection indicators, and unsafe auth/crypto handling.
Dependency vulnerability matching (CVE-aware): Detects known CVEs in dependencies by matching lockfiles against CVE databases, then reports remediation guidance.
Agent threat modeling for tool calls: Audits tool-call patterns for dangerous commands and unsafe agent behavior, with severity and CWE classification.

Quick Start

Use the agent-security-scanning skill to scan AI-generated source code and its dependencies for OWASP/LLM-specific issues and generate a remediation report.

Frequently Asked Questions about agent-security-scanning

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I detect prompt injection vulnerabilities in AI-generated code?▼

Detect prompt injection vulnerabilities by modeling agent-specific threats and scanning tool-call patterns for data exfiltration. Security scanning applies OWASP Top 10 for LLM Applications to flag unsafe agent behaviors before production deployment.

What is the best way to scan dependencies for known CVEs before deployment?▼

The best way to scan dependencies for known CVEs is matching lockfiles against CVE databases. Dependency vulnerability matching detects known issues and generates remediation guidance for pre-commit and CI security gates.

How does agent threat modeling identify data exfiltration paths?▼

Agent threat modeling identifies data exfiltration paths by auditing tool-call patterns for dangerous commands and unsafe behaviors. It applies severity ratings and CWE classification to highlight risks in agent tool execution.

Can I use static analysis to find command injection indicators in AI-generated source code?▼

Yes, static code scanning identifies command injection indicators, unsanitized dynamic SQL, XSS, and SSRF patterns in AI-generated source code. It targets vulnerability patterns tied to agent-generated behavior.

Does agent-security-scanning require specific lockfile formats to run dependency audits?▼

Dependency audits require lockfiles to match against CVE databases for known vulnerabilities. Security scanning requires lockfiles to detect known CVEs and generate concrete remediation recommendations for each finding.