ctf-ai-ml

Solves authorized CTF challenges involving AI, ML models, adversarial learning, and LLM security.

1|Updated Aug 3, 2026
One-click install
npx skills add https://github.com/hanmujun/hanmujun-agent-public --skill ctf-ai-ml-hanmujun
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ctf-ai-ml
Source: https://github.com/hanmujun/hanmujun-agent-public/tree/main/.agents/skills/ctf-ai-ml
Command: npx skills add https://github.com/hanmujun/hanmujun-agent-public --skill ctf-ai-ml-hanmujun

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It provides a structured entry point for tackling authorized CTF challenges in the AI/ML category, where model exploitation, adversarial examples, and LLM security issues require specialized techniques that general-purpose solving approaches miss. ## Core Features & Use Cases - AI/ML Challenge Routing: Directs the agent to the full technical manual at skill-library/ctf-ai-ml/SKILL.md before attempting solutions. - Evidence-First Workflow: Requires reading relevant references and saving experiments and evidence rather than submitting flags directly. - Use Case: When a CTF task involves a pickle deserialization flaw in an ML model or a prompt injection vulnerability in an LLM service, load this skill to access the dedicated methodology and reference material for that challenge type. ## Quick Start Load the ctf-ai-ml skill and read the full manual in the skill library before analyzing the current AI or LLM security challenge.

Frequently Asked Questions about ctf-ai-ml

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

FAQPage Schema
How do I solve AI and ML CTF challenges?▼

Load this skill as the entry point, then read the full technical manual at skill-library/ctf-ai-ml/SKILL.md along with the references relevant to the current challenge. Save your experiments and evidence before attempting any flag submission.

What challenge types does the ctf-ai-ml skill cover?▼

It covers authorized CTF challenges involving AI systems, machine learning models, adversarial learning techniques, and LLM security issues. The skill acts as a bridge to a detailed manual containing the specific methodologies for each challenge type.

Can I submit flags directly when using this skill?▼

No. The skill explicitly instructs you to read the full manual and required references first, then save experiments and evidence. Direct flag submission without this verification workflow is discouraged.

Where is the full documentation for AI/ML CTF techniques?▼

The complete technical manual is located at skill-library/ctf-ai-ml/SKILL.md. This bridge skill only routes you there; the detailed references and procedures live in the skill library directory.

Is this skill limited to authorized CTF environments?▼

Yes. The skill description explicitly scopes it to authorized CTF challenges, meaning competition platforms or environments where the organizer has granted permission for security testing.