Twitter/X Launch Thread: Anthropic-Cybersecurity-Skills v1.0.0

Post timing: Tuesday, 10:00 AM EST (peak developer engagement window) Total character budget: Each tweet <= 280 characters

Tweet 1 / 6 (Hook)

We just open-sourced 734 structured cybersecurity skills for AI agents.

Covering 37 subdomains from malware analysis to zero trust -- each with runnable scripts, reference docs, and MITRE ATT&CK mappings.

[IMAGE: repo structure screenshot showing skill folders]

Thread on what's inside:

Characters: ~272 Image: Screenshot of repo root showing skill directories in a terminal or GitHub file browser.

Tweet 2 / 6 (The Problem)

The problem: AI agents fail at security tasks because they lack structured operational knowledge.

Ask an agent to write a Sigma rule, triage a phishing email, or analyze a memory dump -- you get generic advice, not actionable procedures.

These skills fix that.

Characters: ~270

Tweet 3 / 6 (What's Inside -- Domain Breakdown)

734 skills across 37 subdomains:

- Cloud Security: 60
- Threat Hunting: 55
- Threat Intelligence: 50
- Web App Security: 42
- Network Security: 40
- Malware Analysis: 39
- Digital Forensics: 37
- SOC Operations: 33
- IAM: 35
- + 28 more subdomains

All Apache-2.0 licensed.

Characters: ~275

Tweet 4 / 6 (Skill Anatomy)

Each skill includes:

- SKILL.md with workflow steps, key concepts, and prerequisites
- /scripts with runnable Python (1000+ scripts total)
- /references with tool-specific docs
- MITRE ATT&CK tags (127+ unique technique IDs mapped)
- Domain/subdomain metadata for filtering

No placeholders. Real implementations.

Characters: ~278

Tweet 5 / 6 (Quick Start)

Get started in 30 seconds:

git clone https://github.com/anthropics/anthropic-cybersecurity-skills
cd skills/detecting-mimikatz-execution-patterns
cat SKILL.md

Feed any skill to your AI agent as context. Works with Claude, GPT, or any LLM-based workflow.

github.com/anthropics/anthropic-cybersecurity-skills

Characters: ~279 Note: Verify actual GitHub URL before posting. Placeholder org used.

Tweet 6 / 6 (CTA)

If you work in security or build AI agents for security workflows, this is for you.

Star the repo. Open a PR. Add skills we missed.

cc @DanielMiessler @likethecoins @marcusjcarey -- would love your take.

#cybersecurity #opensource #AISecurity #infosec #MITREATTaCK

Characters: ~274

Tagging Strategy

Account	Followers	Relevance
@DanielMiessler	139K	AI + security thought leader, runs Unsupervised Learning, built fabric (AI framework)
@likethecoins	55K	Cybersecurity practitioner with engaged audience
@marcusjcarey	53K	Author of Tribe of Hackers, pen testing / IR background

All three accounts verified active on X as of 2025-2026.

Thread Scheduling Notes

Best day: Tuesday (highest X engagement for technical content)
Best time: 10:00 AM EST / 7:00 AM PST
Alt window: Wednesday 9:00 AM - 12:00 PM EST
Post tweet 1, then replies 2-6 in rapid succession (within 2 minutes) to keep the thread connected
Pin tweet 1 to profile after posting
Quote-tweet tweet 1 from any organizational account for amplification

3.2 KiB Raw Blame History