Anthropic-Cybersecurity-Skills/launch/hacker-news.md

Hacker News - Show HN Post

Title

Show HN: 611+ Cybersecurity Skills for AI Agents (agentskills.io open standard)

Body

I built an open-source database of 611+ cybersecurity skills that AI agents can use to perform real security work -- from malware analysis with Volatility to cloud pen-testing with Pacu.

Each skill follows a structured format (YAML frontmatter + Markdown body) with:

• When to use (and when NOT to)
• Prerequisites and tool requirements
• Step-by-step workflows with real commands, not pseudocode
• References to real standards (NIST, MITRE ATT&CK, CIS)
• Practitioner helper scripts and report templates

The 611 skills cover 24 subdomains: cloud security, threat intelligence, web app security, threat hunting, malware analysis, digital forensics, SOC operations, network security, IAM, OT/ICS security, API security, container security, vulnerability management, red teaming, incident response, penetration testing, zero trust, phishing defense, endpoint security, DevSecOps, cryptography, mobile security, ransomware defense, and compliance/governance.

Why I built this: AI coding agents (Claude Code, Cursor, Copilot) are great at software engineering but have no structured cybersecurity knowledge. When you ask them to analyze a memory dump or triage a SIEM alert, they give generic advice instead of the precise Volatility plugin sequence or Splunk SPL query a practitioner would use.

The skills use "progressive disclosure" -- the frontmatter tells the agent WHEN to activate a skill, and the full body provides the HOW with exact commands, flags, and decision trees.

Format follows the agentskills.io open standard so any agent framework can consume them.

Repo: https://github.com/mukul975/Anthropic-Cybersecurity-Skills

Tech details:

• Each skill has SKILL.md + references/ + scripts/ + assets/
• Skills are tool-specific (not "use a scanner" but "use Nessus with these plugin families")
• Real CVE references, real MITRE ATT&CK technique IDs
• MIT licensed

Looking for contributors, especially practitioners who want to improve existing skills or add new ones in underrepresented areas (mobile security, OT/ICS, compliance).

Timing

Post at 6:00 AM Pacific (9:00 AM Eastern) for maximum HN visibility. Tuesday through Thursday are optimal days.

Engagement Rules

1. Respond within 1 hour of posting. Early engagement signals to the HN algorithm that the post is active and worth ranking higher.
2. Be technical, not promotional. Answer questions with specific technical details. Reference actual skill files, tool commands, and MITRE technique IDs.
3. No vote requests. Never ask anyone to upvote. This violates HN guidelines and can get the post flagged or penalized.
4. Engage authentically. If someone raises a valid criticism (e.g., "these skills are too shallow for real practitioners"), acknowledge it and explain the contribution model.
5. Link to specific skills when answering questions. For example: "Here's the actual Volatility skill that covers that -- [link to SKILL.md]"
6. Avoid marketing language. No "revolutionary," "game-changing," or "disrupting." Stick to factual descriptions of what the repo contains.
7. Be prepared for skepticism. HN users will question whether AI agents can actually do security work. Have concrete examples ready showing how an agent uses a skill file.