Skip to content
Get packs

AI/LLM App Attack Testing

Guides an agent through security tests of LLM apps, agents, RAG systems and MCP plugins.

Source Download skill
Repository
AIPentest/CyberStrikeAI
Stars
8,323
Licence
Apache-2.0
Files
1 file

What it does

This skill is a set of notes that tells an AI agent how to probe applications built on large language models for security weaknesses. It lists attack categories such as prompt injection (including hidden instructions in documents, web pages or images), abuse of agent tools (for example, making a tool fetch internal cloud metadata or run commands), RAG data poisoning, MCP plugin supply-chain risks, and unsafe loading of PyTorch model files via pickle. It also says a finding should only be recorded once a tool side effect is actually confirmed, such as an out-of-band callback or a file being read.

Great for

  • Penetration testers assessing LLM-powered applications or agents they are authorised to test
  • Security students learning how AI application attack surfaces work
  • Red teams reviewing MCP plugins or RAG pipelines before deployment

Skip it if

  • Testing systems you do not own or have written permission to test, since the content describes live exploitation techniques and could cause harm or legal trouble
  • Developers looking for defensive hardening guidance, because the file is written from an attacker's perspective and gives no mitigations

Try asking

  • “Test this chatbot's agent for prompt injection that leads to tool misuse, with my written authorisation”
  • “Check whether our RAG pipeline can be poisoned through uploaded documents”
  • “Review this MCP plugin for supply-chain risks and SSRF through its fetch tool”
  • “Look at this project's model loading code for torch.load pickle RCE risk”
Step-by-step guide: how to use AI/LLM App Attack TestingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    ai-llm-app-attack.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip ai-llm-app-attack.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip ai-llm-app-attack.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Only use this on targets you are explicitly authorised to test, and keep a written scope.
  • Provide the agent with the application's endpoints or tool list so it can test concrete surfaces rather than guessing.
  • Treat the skill as a checklist; the agent still needs to run its own tests against the target, which may involve live network traffic.

What's inside

  • SKILL.md

Redistributed unmodified from AIPentest/CyberStrikeAI under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.