Skip to content
  1. Home
  2. Guides and documents
  3. OWASP Top 10 for Large Language Model Applications

OWASP Top 10 for Large Language Model Applications

Source repository for OWASP's top ten risks in LLM applications

What it is

The official source of the OWASP Top 10 for LLM Applications, a community awareness document listing the main risk categories, such as prompt injection and sensitive information disclosure, with examples and mitigations. The 2026 release is kept as Markdown alongside the 2025 edition.

Good for

  • Review an LLM app against the ten risks
  • Cite a risk category in a threat model
  • Compare the 2025 and 2026 lists
Worth knowingAn awareness document, not a testing procedure; an older OWASP repo still exists as a legacy entry point.

Sources

The summary is ours, written from the README. Spotted an error? See how we work.

More for security & quality

See the topic →
  • OpenAI Cookbook: Evaluation Examples

    openai/openai-cookbook

    Worked examples for evaluating prompts, RAG and SQL generation

    GuideGitHub stars: 77k
  • Trail of Bits Claude Code Config

    trailofbits/claude-code-config

    Security firm's documented Claude Code defaults and sandboxing setup

    GuideGitHub stars: 2.1k
  • Awesome MCP Security

    Puliczek/awesome-mcp-security

    Reading list on Model Context Protocol security risks and tools

    GuideGitHub stars: 747
  • Awesome Prompt Injection

    Joe-B-Security/awesome-prompt-injection

    Reading list on prompt injection: articles, papers, tools and CTFs

    GuideGitHub stars: 721