# Anthropic-Cybersecurity-Skills v1.3.0 — 817 security skills across 6 frameworks

> 55 new agent skills for AI security, supply chain, and hardware/firmware land in the most-starred open cybersecurity skills pack.

- Org: Mahipal Jangra
- Date: 2026-06-22
- Category: repo, security
- Source: https://github.com/mukul975/Anthropic-Cybersecurity-Skills/releases/tag/v1.3.0
- Canonical page: https://ai-tldr.dev/releases/mukul975-cybersecurity-skills-v1-3/

## In short

Mahipal Jangra's open Anthropic-Cybersecurity-Skills library jumps from 762 to 817 agent skills in v1.3.0, adding AI Security, Supply Chain, and Hardware/Firmware domains plus MITRE F3 as a sixth framework mapping.

## What it is

Anthropic-Cybersecurity-Skills v1.3.0 is the latest release of a community pack of 817 structured security skills for AI coding agents, by maintainer Mahipal Jangra. It is not an Anthropic product — the name reflects the file format. Each skill is a Markdown playbook with metadata, mapped to industry frameworks. The pack works with Claude Code, GitHub Copilot, Cursor, Codex CLI, Gemini CLI, Windsurf, Cline, Continue, and other agentskills.io-compatible tools.

## How it works

The v1.3.0 cut grows the library from 762 to 817 skills and introduces three new domains: AI Security (LLM red-teaming, prompt injection, RAG poisoning), Supply Chain Security (SBOM generation, dependency confusion, malicious-package detection), and Hardware/Firmware Security (UEFI audits, Secure Boot analysis, TPM attestation). Every skill is revalidated against MITRE ATT&CK v19.1 and now also tagged with MITRE Fight Fraud (F3 v1.1) — the sixth framework mapping alongside ATT&CK, NIST CSF 2.0, ATLAS, D3FEND, and NIST AI RMF.

## Why it matters

Skill packs decide what a coding agent can do without re-prompting. With 18.8k stars and pickup across 20+ platforms, this pack is becoming a de facto starting kit for any team wiring agents into incident response, threat hunting, or red-team workflows. The added AI Security domain in v1.3.0 specifically targets LLM-on-LLM attacks — a category that grew straight into production this quarter.

## Who it's for

Security engineers and SOC teams building agent-driven workflows for threat hunting, incident response, or red-team automation.

## Try it

npx skills add mukul975/Anthropic-Cybersecurity-Skills

## Links

- [Release v1.3.0](https://github.com/mukul975/Anthropic-Cybersecurity-Skills/releases/tag/v1.3.0)
- [Repository](https://github.com/mukul975/Anthropic-Cybersecurity-Skills)
- [Project page](https://mahipal.engineer/Anthropic-Cybersecurity-Skills/)