Anthropic-Cybersecurity-Skills
Anthropic-Cybersecurity-Skills
- 评测生成时间(北京时间)
- 本报告引擎
- v3.16.0
- 当前引擎
- v3.17.3
本报告与当前引擎使用不同规则;原分数不会自动更新,不同版本的分数不宜直接对比。
进入后确认来源与额度,提交才会创建任务。
代码版本未记录或记录无效,无法核对本报告的 commit 与送评文本;不会用当前项目代码回填历史记录。
综合采用结论
检测到关键风险,修复前不建议接入
- 基础评测完成+25/25确定性评分与静态安全扫描已完成
- README 有效证据+25/2519,793 个去重后的有效字符
- 独立证据来源+20/205 类非重复证据,重复文件不叠加
- 仓库元数据+10/10已取得仓库状态与采用数据
- 活跃记录+5/5已取得最近提交时间
- AI 复核+15/15已完成结构化 AI 证据复核
本报告评了什么
综合分用于静态采用筛选,不是任务成功率,也不代表使用 Skill 后的能力提升。
- 静态检查范围
- 记录 25 个文件 · 214,936 字符
- 只检查采集到的受限文本与已知风险模式;未读取的脚本、依赖和运行环境不在结论内。
- AI 内容复核
- 已使用
- 复核送评文本的说明和设计,不是安装测试或实际任务执行。记录缺失时不推断已使用或未使用。
- 真实任务效果
- 本报告未提供实测证据
- 没有使用/不使用 Skill 的任务对照、重复运行或耗时与 token 成本结果。采用前仍需在隔离环境验证自己的任务。
脚本引用覆盖
脚本引用覆盖记录不可得;不推断没有引用或已经检查。
磁盘镜像采集与校验流程
SKILL.md 的 Workflow 为单一任务的连续六步处理,含识别、写保护、采集、校验、报告,符合 flow 类型
左右滑动查看完整图示
- • Step 1 使用 blockdev --setro 启用软件写保护
- • Step 4 使用 dcfldd 采集并写入 hashlog
- • Step 5 用 diff 比较源与镜像 SHA-256
- 通过: Agent Skills 格式校验 24/24 个通过
- 通过: 24 个有效 Skill 含可核实的指令步骤或示例
- 质量评审选取 skills/acquiring-disk-image-with-dd-and-dcfldd/SKILL.md;其余 23 个仅做格式扫描
- SKILL.md 标题为 'Acquiring Disk Image with dd and dcfldd',domain 为 cybersecurity,subdomain 为 digital-forensics
- Step 3 给出 dd if=/dev/sdb of=... bs=4096 conv=noerror,sync status=progress 命令
- Step 4 给出 dcfldd 参数 hash=sha256,md5、hashwindow=1G、hashlog、split=2G、vf 验证
- Step 5 用 diff 比较源与镜像 SHA-256,并再次哈希源设备确认未变更
- README 声明 754 skills、26 domains、Apache-2.0 许可,Quick start 提供 npx skills add 与 git clone
- 有效 README
- 安装或接入步骤
- 可执行示例
- 输入、参数或工具说明
- 通过: Agent Skills 格式校验 24/24 个通过
- 发现提示词或系统信息提取意图
- 发现疑似 LLM API Key
- skills/acquiring-disk-image-with-dd-and-dcfldd/SKILL.md:description 未清楚说明何时使用该 Skill
- skills/analyzing-active-directory-acl-abuse/SKILL.md:description 未清楚说明何时使用该 Skill
- skills/analyzing-android-malware-with-apktool/SKILL.md:description 未清楚说明何时使用该 Skill
需要按取证规范制作磁盘位对位镜像的 DFIR 人员、教学或演练中演示 dd/dcfldd 与哈希校验流程、为 AI Agent 提供结构化磁盘采集步骤
需要完整错误处理与排障手册的生产环境、非 Linux 取证工作站环境
也有自己的公开项目?先免登录查已有报告;没有报告时,可保留地址继续登录评测。
查找我项目的报告 →prompt-extractionREADME.md:226high confidence| Exfiltration | TA0010 | Strong | DNS exfiltration, DLP controls, data loss detection |修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
llm-api-keyskills/acquiring-disk-image-with-dd-and-dcfldd/SKILL.md:2high confidencename: acquiring-disk-***redacted***修复:立即轮换密钥,并改用环境变量或 Secret Manager 注入。
prompt-extractionskills/analyzing-apt-group-with-mitre-navigator/SKILL.md:256high confidence"ex***on", "impact",修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
prompt-extractionskills/analyzing-cloud-storage-access-patterns/SKILL.md:5high confidencecalls (GetObject spikes), and potential data exfiltration using statistical baselines and time-series anomaly detection.修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
prompt-extractionskills/analyzing-command-and-control-communication/SKILL.md:75high confidenceEmail: SMTP/IMAP for C2 commands and data exfiltration修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
prompt-extractionskills/analyzing-command-and-control-communication/SKILL.md:381high confidence0x04 - Upload Exfiltrate file to C2修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
prompt-extractionskills/analyzing-cyber-kill-chain/SKILL.md:71high confidence- Indicators: Data staging/exfiltration, lateral movement, ransomware execution, destructive activity修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
prompt-extractionskills/analyzing-cyber-kill-chain/SKILL.md:95high confidence- Actions on Objectives → Exfiltration (TA0010), Impact (TA0040)修复:移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
- 01移除提示词提取逻辑,并增加敏感上下文不可输出的边界说明。
- 02立即轮换密钥,并改用环境变量或 Secret Manager 注入。
- 03修复 skills/acquiring-disk-image-with-dd-and-dcfldd/SKILL.md:description 未清楚说明何时使用该 Skill
- 04修复 skills/analyzing-active-directory-acl-abuse/SKILL.md:description 未清楚说明何时使用该 Skill
- 05修复 skills/analyzing-android-malware-with-apktool/SKILL.md:description 未清楚说明何时使用该 Skill
- 06修复 skills/analyzing-apt-group-with-mitre-navigator/SKILL.md:description 未清楚说明何时使用该 Skill
方法、证据与局限展开收起
GitHub Repository API
25 个文件 · 214,936 字符
v3.16.0 · AI 内容复核:已使用(deepseek-flash)
- 静态评测不会安装或执行项目代码
- 安全扫描基于高信号文件与已知模式,不能替代人工审计
- 流行度只反映采用程度,不代表安全或工程质量
- 发现 24 个 Skill;质量复核只评审 skills/acquiring-disk-image-with-dd-and-dcfldd/SKILL.md,其余 23 个仅做格式扫描
30 天热度趋势
README
Anthropic Cybersecurity Skills
The largest open-source cybersecurity skills library for AI agents
754 production-grade cybersecurity skills · 26 security domains · 5 framework mappings · 26+ AI platforms
Get Started · What's Inside · Frameworks · Platforms · Contributing
⚠️ Community Project — This is an independent, community-created project. Not affiliated with Anthropic PBC.
Give any AI agent the security skills of a senior analyst
A junior analyst knows which Volatility3 plugin to run on a suspicious memory dump, which Sigma rules catch Kerberoasting, and how to scope a cloud breach across three providers. Your AI agent doesn't — unless you give it these skills.
This repo contains 754 structured cybersecurity skills spanning 26 security domains, each following the agentskills.io open standard. Every skill is mapped to five industry frameworks — MITRE ATT&CK, NIST CSF 2.0, MITRE ATLAS, MITRE D3FEND, and NIST AI RMF — making this the only open-source skills library with unified cross-framework coverage. Clone it, point your agent at it, and your next security investigation gets expert-level guidance in seconds.
Five frameworks, one skill library
No other open-source skills library maps every skill to all five frameworks. One skill, five compliance checkboxes.
| Framework | Version | Scope in this repo | What it maps |
|---|---|---|---|
| MITRE ATT&CK | v18 | 14 tactics · 200+ techniques | Adversary behaviors and TTPs |
| NIST CSF 2.0 | 2.0 | 6 functions · 22 categories | Organizational security posture |
| MITRE ATLAS | v5.4 | 16 tactics · 84 techniques | AI/ML adversarial threats |
| MITRE D3FEND | v1.3 | 7 categories · 267 techniques | Defensive countermeasures |
| NIST AI RMF | 1.0 | 4 functions · 72 subcategories | AI risk management |
Example — a single skill maps across all five:
| Skill | ATT&CK | NIST CSF | ATLAS | D3FEND | AI RMF |
|---|---|---|---|---|---|
analyzing-network-traffic-of-malware | T1071 | DE.CM | AML.T0047 | D3-NTA | MEASURE-2.6 |
Quick start
# Option 1: npx (recommended)
npx skills add mukul975/Anthropic-Cybersecurity-Skills
# Option 2: Git clone
git clone https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git
cd Anthropic-Cybersecurity-Skills
Works immediately with Claude Code, GitHub Copilot, OpenAI Codex CLI, Cursor, Gemini CLI, and any agentskills.io-compatible platform.
🚀 Try it on the Playground
Experience Casky.ai hands-on — no setup required.
→ Launch Playground on Casky.ai
The playground lets you:
- Run live cybersecurity skill exercises against real targets
- See AI agents execute structured skills in real time
- Explore MITRE ATT&CK mapped workflows interactively
- Test threat hunting, DFIR, and penetration testing scenarios
No installation. No configuration. Just open and start.
Why this exists
The cybersecurity workforce gap hit 4.8 million unfilled roles globally in 2024 (ISC2). AI agents can help close that gap — but only if they have structured domain knowledge to work from. Today's agents can write code and search the web, but they lack the practitioner playbooks that turn a generic LLM into a capable security analyst.
Existing security tool repos give you wordlists, payloads, or exploit code. None of them give an AI agent the structured decision-making workflow a senior analyst follows: when to use each technique, what prerequisites to check, how to execute step-by-step, and how to verify results. That is the gap this project fills.
Anthropic Cybersecurity Skills is not a collection of scripts or checklists. It is an AI-native knowledge base built from the ground up for the agentskills.io standard — YAML frontmatter for sub-second discovery, structured Markdown for step-by-step execution, and reference files for deep technical context. Every skill encodes real practitioner workflows, not generated summaries.
What's inside — 26 security domains
| Domain | Skills | Key capabilities |
|---|---|---|
| Cloud Security | 60 | AWS, Azure, GCP hardening · CSPM · cloud forensics |
| Threat Hunting | 55 | Hypothesis-driven hunts · LOTL detection · behavioral analytics |
| Threat Intelligence | 50 | STIX/TAXII · MISP · feed integration · actor profiling |
| Web Application Security | 42 | OWASP Top 10 · SQLi · XSS · SSRF · deserialization |
| Network Security | 40 | IDS/IPS · firewall rules · VLAN segmentation · traffic analysis |
| Malware Analysis | 39 | Static/dynamic analysis · reverse engineering · sandboxing |
| Digital Forensics | 37 | Disk imaging · memory forensics · timeline reconstruction |
| Security Operations | 36 | SIEM correlation · log analysis · alert triage |
| Identity & Access Management | 35 | IAM policies · PAM · zero trust identity · Okta · SailPoint |
| SOC Operations | 33 | Playbooks · escalation workflows · metrics · tabletop exercises |
| Container Security | 30 | K8s RBAC · image scanning · Falco · container forensics |
| OT/ICS Security | 28 | Modbus · DNP3 · IEC 62443 · historian defense · SCADA |
| API Security | 28 | GraphQL · REST · OWASP API Top 10 · WAF bypass |
| Vulnerability Management | 25 | Nessus · scanning workflows · patch prioritization · CVSS |
| Incident Response | 25 | Breach containment · ransomware response · IR playbooks |
| Red Teaming | 24 | Full-scope engagements · AD attacks · phishing simulation |
| Penetration Testing | 23 | Network · web · cloud · mobile · wireless pentesting |
| Endpoint Security | 17 | EDR · LOTL detection · fileless malware · persistence hunting |
| DevSecOps | 17 | CI/CD security · code signing · Terraform auditing |
| Phishing Defense | 16 | Email authentication · BEC detection · phishing IR |
| Cryptography | 14 | TLS · Ed25519 · certificate transparency · key management |
| Zero Trust Architecture | 13 | BeyondCorp · CISA maturity model · microsegmentation |
| Mobile Security | 12 | Android/iOS analysis · mobile pentesting · MDM forensics |
| Ransomware Defense | 7 | Precursor detection · response · recovery · encryption analysis |
| Compliance & Governance | 5 | CIS benchmarks · SOC 2 · regulatory frameworks |
| Deception Technology | 2 | Honeytokens · breach detection canaries |
How AI agents use these skills
Each skill costs ~30 tokens to scan (frontmatter only) and 500–2,000 tokens to fully load (complete workflow). This progressive disclosure architecture lets agents search all 754 skills in a single pass without blowing context windows.
User prompt: "Analyze this memory dump for signs of credential theft"
Agent's internal process:
1. Scans 754 skill frontmatters (~30 tokens each)
→ identifies 12 relevant skills by matching tags, description, domain
2. Loads top 3 matches:
• performing-memory-forensics-with-volatility3
• hunting-for-credential-dumping-lsass
• analyzing-windows-event-logs-for-credential-access
3. Executes the structured Workflow section step-by-step
→ runs Volatility3 plugins, checks LSASS access patterns,
correlates with event log evidence
4. Validates results using the Verification section
→ confirms IOCs, maps findings to ATT&CK T1003 (Credential Dumping)
Without these skills, the agent guesses at tool commands and misses critical steps. With them, it follows the same playbook a senior DFIR analyst would use.
Skill anatomy
Every skill follows a consistent directory structure:
skills/performing-memory-forensics-with-volatility3/
├── SKILL.md ← Skill definition (YAML frontmatter + Markdown body)
├── references/
│ ├── standards.md ← MITRE ATT&CK, ATLAS, D3FEND, NIST mappings
│ └── workflows.md ← Deep technical procedure reference
├── scripts/
│ └── process.py ← Working helper scripts
└── assets/
└── template.md ← Filled-in checklists and report templates
YAML frontmatter (real example)
---
name: performing-memory-forensics-with-volatility3
description: >-
Analyze memory dumps to extract running processes, network connections,
injected code, and malware artifacts using the Volatility3 framework.
domain: cybersecurity
subdomain: digital-forensics
tags: [forensics, memory-analysis, volatility3, incident-response, dfir]
atlas_techniques: [AML.T0047]
d3fend_techniques: [D3-MA, D3-PSMD]
nist_ai_rmf: [MEASURE-2.6]
nist_csf: [DE.CM-01, RS.AN-03]
version: "1.2"
author: mukul975
license: Apache-2.0
---
Markdown body sections
## When to Use
Trigger conditions — when should an AI agent activate this skill?
## Prerequisites
Required tools, access levels, and environment setup.
## Workflow
Step-by-step execution guide with specific commands and decision points.
## Verification
How to confirm the skill was executed successfully.
Frontmatter fields: name (kebab-case, 1–64 chars), description (keyword-rich for agent discovery), domain, subdomain, tags, atlas_techniques (MITRE ATLAS IDs), d3fend_techniques (MITRE D3FEND IDs), nist_ai_rmf (NIST AI RMF references), nist_csf (NIST CSF 2.0 categories). MITRE ATT&CK technique mappings are documented in each skill's references/standards.md file and in the ATT&CK Navigator layer included with releases.
📊 MITRE ATT&CK Enterprise coverage — all 14 tactics
| Tactic | ID | Coverage | Key skills |
|---|---|---|---|
| Reconnaissance | TA0043 | Strong | OSINT, subdomain enumeration, DNS recon |
| Resource Development | TA0042 | Moderate | Phishing infrastructure, C2 setup detection |
| Initial Access | TA0001 | Strong | Phishing simulation, exploit detection, forced browsing |
| Execution | TA0002 | Strong | PowerShell analysis, fileless malware, script block logging |
| Persistence | TA0003 | Strong | Scheduled tasks, registry, service accounts, LOTL |
| Privilege Escalation | TA0004 | Strong | Kerberoasting, AD attacks, cloud privilege escalation |
| Defense Evasion | TA0005 | Strong | Obfuscation, rootkit analysis, evasion detection |
| Credential Access | TA0006 | Strong | Mimikatz detection, pass-the-hash, credential dumping |
| Discovery | TA0007 | Moderate | BloodHound, AD enumeration, network scanning |
| Lateral Movement | TA0008 | Strong | SMB exploits, lateral movement detection with Splunk |
| Collection | TA0009 | Moderate | Email forensics, data staging detection |
| Command and Control | TA0011 | Strong | C2 beaconing, DNS tunneling, Cobalt Strike analysis |
| Exfiltration | TA0010 | Strong | DNS exfiltration, DLP controls, data loss detection |
| Impact | TA0040 | Strong | Ransomware defense, encryption analysis, recovery |
An ATT&CK Navigator layer file is included in the v1.0.0 release assets for visual coverage mapping.
Note: ATT&CK v19 lands April 28, 2026 — splitting Defense Evasion (TA0005) into two new tactics: Stealth and Impair Defenses. Skill mappings will be updated in a forthcoming release.
📊 NIST CSF 2.0 alignment — all 6 functions
| Function | Skills | Examples |
|---|---|---|
| Govern (GV) | 30+ | Risk strategy, policy frameworks, roles & responsibilities |
| Identify (ID) | 120+ | Asset discovery, threat landscape assessment, risk analysis |
| Protect (PR) | 150+ | IAM hardening, WAF rules, zero trust, encryption |
| Detect (DE) | 200+ | Threat hunting, SIEM correlation, anomaly detection |
| Respond (RS) | 160+ | Incident response, forensics, breach containment |
| Recover (RC) | 40+ | Ransomware recovery, BCP, disaster recovery |
NIST CSF 2.0 (February 2024) added the Govern function and expanded scope from critical infrastructure to all organizations. Skill mappings align to all 22 categories and reference 106 subcategories.
📊 Framework deep dive — ATLAS, D3FEND, AI RMF
MITRE ATLAS v5.4 — AI/ML adversarial threats
ATLAS maps adversarial tactics, techniques, and case studies specific to AI and machine learning systems. Version 5.4 covers 16 tactics and 84 techniques including agentic AI attack vectors added in late 2025: AI agent context poisoning, tool invocation abuse, MCP server compromises, and malicious agent deployment. Skills mapped to ATLAS help agents identify and defend against threats to ML pipelines, model weights, inference APIs, and autonomous workflows.
MITRE D3FEND v1.3 — Defensive countermeasures
D3FEND is an NSA-funded knowledge graph of 267 defensive techniques organized across 7 tactical categories: Model, Harden, Detect, Isolate, Deceive, Evict, and Restore. Built on OWL 2 ontology, it uses a shared Digital Artifact layer to bidirectionally map defensive countermeasures to ATT&CK offensive techniques. Skills tagged with D3FEND identifiers let agents recommend specific countermeasures for detected threats.
NIST AI RMF 1.0 + GenAI Profile (AI 600-1)
The AI Risk Management Framework defines 4 core functions — Govern, Map, Measure, Manage — with 72 subcategories for trustworthy AI development. The GenAI Profile (AI 600-1, July 2024) adds 12 risk categories specific to generative AI, from confabulation and data privacy to prompt injection and supply chain risks. Colorado's AI Act (effective February 2026) provides a legal safe harbor for organizations complying with NIST AI RMF, making these mappings directly relevant to regulatory compliance.
Compatible platforms
AI code assistants Claude Code (Anthropic) · GitHub Copilot (Microsoft) · Cursor · Windsurf · Cline · Aider · Continue · Roo Code · Amazon Q Developer · Tabnine · Sourcegraph Cody · JetBrains AI
CLI agents OpenAI Codex CLI · Gemini CLI (Google)
Autonomous agents Devin · Replit Agent · SWE-agent · OpenHands
Agent frameworks & SDKs LangChain · CrewAI · AutoGen · Semantic Kernel · Haystack · Vercel AI SDK · Any MCP-compatible agent
All platforms that support the agentskills.io standard can load these skills with zero configuration.
What people are saying
"A database of real, organized security skills that any AI agent can plug into and use. Not tutorials. Not blog posts." — Hasan Toor (@hasantoxr), AI/tech creator
"This is not a random collection of security scripts. It's a structured operational knowledge base designed for AI-driven security workflows." — fazal-sec, Medium
Featured in
| Where | Type | Link |
|---|---|---|
| awesome-agent-skills | Awesome List (1,000+ skills index) | VoltAgent/awesome-agent-skills |
| awesome-ai-security | Awesome List (AI security tools) | ottosulin/awesome-ai-security |
| awesome-codex-cli | Awesome List (Codex CLI resources) | RoggeOhta/awesome-codex-cli |
| SkillsLLM | Skills directory & marketplace | skillsllm.com/skill/anthropic-cybersecurity-skills |
| Openflows | Signal analysis & tracking | openflows.org |
| NeverSight skills_feed | Automated skills index | NeverSight/skills_feed |
Star history
Releases
| Version | Date | Highlights |
|---|---|---|
| v1.0.0 | March 11, 2026 | 734 skills · 26 domains · MITRE ATT&CK + NIST CSF 2.0 mapping · ATT&CK Navigator layer |
Skills have continued to grow on main since v1.0.0 — the library now contains 754 skills with 5-framework mapping (MITRE ATLAS, D3FEND, and NIST AI RMF added post-release). Check Releases for the latest tagged version.
Contributing
This project grows through community contributions. Here is how to get involved:
Add a new skill — Domains like Deception Technology (2 skills) and Compliance & Governance (5 skills) need the most help. Follow the template in CONTRIBUTING.md and submit a PR with the title Add skill: your-skill-name.
Improve existing skills — Add framework mappings, fix workflows, update tool references, or contribute scripts and templates.
Report issues — Found an inaccurate procedure or broken script? Open an issue.
Every PR is reviewed for technical accuracy and agentskills.io standard compliance within 48 hours. Check good first issues for a starting point.
This project follows the Contributor Covenant. By participating, you agree to uphold this code.
Community
💬 Discussions — Questions, ideas, and roadmap conversations 🐛 Issues — Bug reports and feature requests 🔒 Security Policy — Responsible disclosure process (48-hour acknowledgment)
Citation
If you use this project in research or publications:
@software{anthropic_cybersecurity_skills,
author = {Jangra, Mahipal},
title = {Anthropic Cybersecurity Skills},
year = {2026},
url = {https://github.com/mukul975/Anthropic-Cybersecurity-Skills},
license = {Apache-2.0},
note = {754 structured cybersecurity skills for AI agents,
mapped to MITRE ATT\&CK, NIST CSF 2.0, MITRE ATLAS,
MITRE D3FEND, and NIST AI RMF}
}
License
This project is licensed under the Apache License 2.0. You are free to use, modify, and distribute these skills in both personal and commercial projects.
If this project helps your security work, consider giving it a ⭐
⭐ Star · 🍴 Fork · 💬 Discuss · 📝 Contribute
Community project by @mukul975. Not affiliated with Anthropic PBC.