kali-claw is a self-evolving security agent from the OpenClaw project. It operates 24/7 on Kali Linux, systematically learning and practicing penetration testing across 49 security domains.
Overview
An AI-powered penetration testing agent built on Kali Linux, mastering all 518 security tools through continuous self-directed learning. is a self-evolving security agent from the OpenClaw project. It operates 24/7 on Kali Linux, systematically learning and practicing penetration testing across . It thinks like a hacker — first principles, divergent thinking, assume breach — and acts like a senior engineer: direct, hands-on, results-driven. : (2026-09-09) — （139 技能三轴验证；语法错误块 102→0；排名报告上线；skill-lint 0/0/139）。详见 RELEASE-v0.3.5.md。
README
kali-claw
An AI-powered penetration testing agent built on Kali Linux, mastering all 518 security tools through continuous self-directed learning.
kali-claw is a self-evolving security agent from the OpenClaw project. It operates 24/7 on Kali Linux, systematically learning and practicing penetration testing across 139 security domains. It thinks like a hacker — first principles, divergent thinking, assume breach — and acts like a senior engineer: direct, hands-on, results-driven.
Current Version: v0.3.5 (2026-09-09) — 全量验证与完整性排序（139 技能三轴验证；语法错误块 102→0；排名报告上线；skill-lint 0/0/139）。详见 RELEASE-v0.3.5.md。
Features
- 139 Security Skill Domains — From OSINT and web exploitation to cloud security + cloud-identity (Entra ID/Okta/Auth0), AI/LLM security (LLM red team + AI agent security + agentic pentest), exploit development, hardware/embedded systems, Bluetooth/RFID/NFC, SCADA/ICS, firmware reverse engineering, VoIP/SIP, database attacks, Active Directory/LDAP attacks, anti-forensics, username-based OSINT dossier generation (Maigret), dark-web intelligence, threat hunting + defensive deception (honeypots) + detection engineering, blockchain/Web3 security, payment security (PCI-DSS), Kubernetes red team, secret-management attack (SAST/secrets), IoT application-layer pentest, physical security testing (locks/badges/USB weapons), quantum/post-quantum/national crypto attacks, deep phishing infrastructure (AiTM/gateway bypass), 5G telecom attack (PFCP/GTP/IMSI catchers/O-RAN), automotive vehicle security (CAN/UDS/key fobs/EV charging), mobile app instrumentation (Frida/Objection/r2frida), cloud-native vulnerability research (CVE methodology/PoC reproduction/nuclei templates), macOS security (SIP/TCC/ESF/Keychain/Apple Silicon), UAV/drone security (MAVLink/PX4/GPS spoofing/DroneID), game anti-cheat bypass (EAC/BattlEye/Vanguard/BYOVD), mainframe security (z/OS/RACF/CICS/DB2/JES2), ICS fieldbus attack (Profibus/DNP3/IEC 61850/IEC 60870-5/EtherCAT/PROFINET), HF/VHF radio attack (ADS-B/AIS/ACARS/POCSAG/APRS), blockchain L2 attack (Lightning Network/Optimism/Arbitrum/zkSync/cross-chain bridges), embedded RTOS security (VxWorks/QNX/FreeRTOS/ThreadX/Zephyr), storage/SAN attack (iSCSI/FC/NFSv4/SMB3/S3/NetApp/Dell EMC/QNAP), hypervisor introspection (VMware ESXi/Hyper-V/KVM/Xen/LibVMI/DRAKVUF/VENOM), satellite/LEO security (Starlink/Iridium/Viasat KA-SAT/DVB-S2/VSAT), AD CS abuse (ESC1-ESC15/PetitPotam/Certifried/Certipy), CI/CD supply chain attack (Jenkins/GitLab CI/GitHub Actions/xz-utils/SolarWinds), PAM privilege attack (CyberArk/BeyondTrust/Delinea/ManageEngine), CSPM/CASB attack (Wiz/Prisma Cloud/Netskope/OPA/Kyverno policy bypass), SASE/SSE attack (Zscaler ZIA/ZPA, Netskope, Cloudflare One, Cisco Umbrella), EU AI Act compliance red team (Article 9 / Annex IV / Notified Body audit), and AI agent supply chain attack (Hugging Face / Pickle RCE / LangChain plugin backdoors / model weight steganography), each with structured payloads, test cases, and learning guides
- Defense Triple Coverage — All 139 SKILLs include Defense Perspective (table format) + Detection Methods (with SIEM rules: Sigma/Splunk SPL/Sysmon/Falco) + Defense Evasion Techniques — 100% coverage
- 12 Hacker Laws — Core behavioral guidelines derived from real-world security philosophy
- Layered Memory System — Daily logs + distilled long-term memory + monthly chronicles for persistent knowledge across sessions
- Heartbeat Task Framework — Automated health checks, security scans, learning progress tracking, and knowledge maintenance
- Fully Reusable — Copy the workspace, change 4 files, and you have a new security agent
Guides
Usage Guide
| Language | File |
|---|---|
| 中文 | GUIDE-OPENCLAW-zh.md |
| English | GUIDE-OPENCLAW-en.md |
Migration Guides
kali-claw is a portable skill package — you can use it with multiple AI agent platforms:
| Platform | 中文 | English |
|---|---|---|
| Hermes Agent | GUIDE-HERMES-zh.md | GUIDE-HERMES-en.md |
| Claude Code | GUIDE-CLAUDECODE-zh.md | GUIDE-CLAUDECODE-en.md |
| OpenAI Codex CLI | GUIDE-CODEX-zh.md | GUIDE-CODEX-en.md |
| OpenCode | GUIDE-OPENCODE-zh.md | GUIDE-OPENCODE-en.md |
New to OpenClaw? Start with the usage guide — it covers everything from installation to your first penetration test.
Quick Start
Prerequisites
- OpenClaw installed and configured
- Recommended: Kali Linux environment — either install OpenClaw directly on Kali Linux, or provide SSH access from your OpenClaw host to a Kali Linux machine so kali-claw can execute security tools remotely
1. Install OpenClaw
npm install -g openclaw@latest
See the official documentation for detailed installation and configuration options.
2. Create a dedicated agent
It is not recommended to use the main agent directly. Create a dedicated agent for this workspace:
openclaw agents add kali-claw --workspace ~/.openclaw/workspace-kali-claw
This creates an isolated agent with its own workspace, auth, and routing. See openclaw agents for full options including --model, --bind, and --non-interactive.
3. Clone this repository into the workspace
cd ~/.openclaw/workspace-kali-claw/
git clone https://github.com/.git .
The agent will automatically read SOUL.md, AGENTS.md, USER.md, and MEMORY.md on startup to initialize its identity and context.
4. Start the agent
openclaw gateway start
Then interact with kali-claw:
> Scan target 192.168.1.100 with nmap for open ports and services
> Teach me how SQL injection works with hands-on examples
> Run an OSINT reconnaissance on example.com
How It Works
Architecture
kali-claw/
├── SOUL.md # Identity + 12 Hacker Laws (the agent's personality)
├── AGENTS.md # Workspace config + session startup sequence
├── IDENTITY.md # Skill tags + personality traits
├── USER.md # Captain (user) profile
├── MEMORY.md # Long-term distilled knowledge
├── TOOLS.md # Tool quick reference + learning progress
├── HEARTBEAT.md # Periodic heartbeat task framework
├── skills/ # 139 security skill domains (v0.2.0.2)
│ ├── api-security/
│ │ ├── SKILL.md # Skill definition + Defense Triple (Defense Perspective + Detection Methods + Defense Evasion)
│ │ ├── payloads.md # Attack payloads
│ │ ├── test-cases.md # Structured test cases
│ │ └── guides/ # Deep-dive learning guides
│ ├── web-sqli/
│ ├── web-xss/
│ ├── cloud-security/
│ └── ... (139 domains total)
├── memory/ # Daily memory logs (YYYY-MM-DD.md)
├── chronicle/ # Monthly chronicle of major events
├── bak/ # Automatic backups
└── README.md # This file
Session Lifecycle
Every time the agent starts a new session:
- Read SOUL.md — Load identity and hacker laws
- Read USER.md — Understand who it’s helping
- Read recent memory — Get context from today and yesterday
- Read MEMORY.md — Load long-term distilled knowledge
The agent wakes up fresh each session but carries continuity through its file-based memory system.
Memory System
| Layer | File | Purpose |
|---|---|---|
| Daily | memory/YYYY-MM-DD.md |
Raw activity logs for the day |
| Long-term | MEMORY.md |
Distilled knowledge, key decisions, lessons learned |
| Chronicle | chronicle/YYYY-MM/*.md |
Monthly record of major events |
Knowledge flows upward: daily logs are regularly distilled into MEMORY.md, and major milestones are recorded in the chronicle.
Security Skills
37 domains organized by the OWASP and MITRE frameworks + 12 additional domains (knowledge operations, AI/LLM, hardware, multi-agent, MCP, bounty hunting, etc.) + 23 expanded domains (exploit development, privilege escalation, payload generation, AV/EDR evasion, DNS attacks, web XXE, file inclusion, CMS attack, steganography, network sniffing & MITM, Bluetooth/RFID/NFC, network tunneling & proxy, firmware reverse, SCADA/ICS security, database attack, VoIP/SIP attack, anti-forensics, pentest reporting, engagement manager, tool mastery, AD/LDAP attack, web deserialization, email protocol attack):
| Domain | Description | Key Topics |
|---|---|---|
api-security |
API security testing | REST/GraphQL testing, auth bypass, rate limiting |
binary-reverse |
Binary analysis & reverse engineering | radare2, exploit development, malware analysis |
cloud-security |
Cloud platform security | AWS/Azure/GCP, IAM, S3 exposure, metadata attacks |
container-security |
Container & K8s security | Docker escape, K8s RBAC, image scanning |
crypto-attacks |
Cryptographic vulnerability testing | Weak algorithms, certificate issues, padding oracle |
digital-forensics |
Digital forensics | Disk analysis, memory forensics, network forensics |
insecure-design |
Insecure design detection | Threat modeling, abuse cases, design patterns |
logging-monitoring |
Logging & monitoring security | Log injection, detection evasion, SIEM bypass |
mobile-security |
Mobile application security | Android/iOS testing, certificate pinning, data leakage |
network-pentest |
Network penetration testing | Scanning, exploitation, lateral movement |
osint |
Open source intelligence | People search, domain recon, data aggregation |
password-attack |
Password attack techniques | Dictionary attacks, hash cracking, rule-based brute force |
post-exploitation |
Post-exploitation operations | Persistence, privilege escalation, data exfiltration |
recon-osint |
Reconnaissance & OSINT | Subdomain enum, port scanning, technology fingerprinting |
security-misconfiguration |
Security misconfiguration detection | Default creds, verbose errors, directory listing |
social-engineering |
Social engineering | Phishing, pretexting, baiting techniques |
social-intelligence |
Social platform intelligence | Reddit/HN/X/YouTube OSINT, community sentiment, target profiling, dark web monitoring |
supply-chain-security |
Software supply chain security | Dependency attacks, CI/CD poisoning, integrity verification |
vulnerability-assessment |
Vulnerability assessment | Automated scanning, manual testing, risk rating |
web-access-control |
Broken access control | IDOR, privilege escalation, forced browsing |
web-auth-bypass |
Authentication bypass | Brute force, session attacks, OAuth flaws |
web-sqli |
SQL injection | Union-based, blind, time-based, double query |
web-ssrf |
Server-Side Request Forgery | Internal scanning, cloud metadata, protocol smuggling |
web-xss |
Cross-Site Scripting | Reflected, stored, DOM-based, CSP bypass |
wifi-pentest |
WiFi penetration testing | WPA cracking, WPS attacks, evil twin |
chronicle |
Chronicle system | Event logging, milestone tracking |
deep-research |
Multi-source intelligence research | CVE deep-dive, threat actor profiling, attack technique investigation, continuous monitoring, intelligence correlation, cited reports |
security-bounty-hunter |
Bug bounty vulnerability hunting | Exploitable vulnerability discovery, PoC development, responsible disclosure reporting |
terminal-ops |
Evidence-first terminal operations | Structured command execution, evidence chain protocol, verified state tracking |
search-first |
Research before exploit | Exploit/tool search workflow, existing solution discovery, decision matrix |
security-review |
Comprehensive security review | OWASP Top 10 checklist, source code audit, configuration review, dependency scanning |
repo-scan |
Cross-stack source code audit | File classification, library detection, module verdicts, security hotspot analysis |
verification-loop |
Multi-phase finding verification | Exploit confirmation, false positive elimination, independent reproduction, evidence documentation |
codebase-onboarding |
Rapid codebase intelligence | 3 scope modes (Targeted/Exploratory/Comprehensive), language tier matrix, confidence scoring, architecture pattern recognition, 100M+ LOC strategy |
knowledge-ops |
Knowledge graph management | Entity extraction, cross-session aggregation, confidence tracking, graph visualization, pattern intelligence |
article-writing |
Security content creation | Pentest reports, CVE disclosures, blog posts, CVSS scoring, sanitization, evidence documentation |
browser-qa |
Automated browser testing | Playwright/Puppeteer, network monitoring, cookie analysis, CSRF detection, XSS payload injection |
data-scraper-agent |
Structured data collection | CVE scraping, exploit DB search, threat intel feeds, GitHub advisories, HTML parsing |
exa-search |
Semantic security research | Context-aware queries, date filtering, domain filtering, full-text extraction, API integration |
ai-fuzzing |
AI-assisted vulnerability discovery | Coverage-guided fuzzing, AFL++/libFuzzer/Honggfuzz, crash triage, Web API fuzzing, protocol fuzzing |
council |
Multi-perspective security analysis | Attack/defense/audit viewpoints, decision matrix, risk assessment, consensus building |
autonomous-loops |
Safe autonomous execution patterns | Sequential pipeline, watch loop, batch processing, learning cycle, scope locks |
continuous-learning |
Engagement knowledge extraction | Pattern detection, confidence scoring, cross-reference linking, memory layering |
docker-patterns |
Docker security testing labs | Vulnerable app labs, network labs, attack chain labs, disposable testing |
safety-guard |
Safety enforcement layer | Scope checking, dangerous command interception, incident response, engagement rules |
bluetooth-rfid-nfc |
Bluetooth/BLE/RFID/NFC attacks | Device discovery, BLE GATT exploitation, MIFARE cracking, NFC cloning |
network-tunneling-proxy |
Network tunneling & proxying | SSH/HTTP tunneling, DNS/ICMP covert tunnels, SOCKS proxy chains, pivoting |
firmware-reverse |
Firmware reverse engineering | Extraction, filesystem analysis, QEMU emulation, backdoor detection |
scada-ics-security |
SCADA/ICS security assessment | Modbus, S7comm, EtherNet/IP, OPC UA, PLC enumeration, honeypots |
database-attack |
Database server attacks | Oracle TNS, Redis/MongoDB unauth, brute-force, stored procedures |
voip-sip-attack |
VoIP/SIP protocol attacks | SIP enumeration, eavesdropping, VLAN hopping, DoS |
anti-forensics |
Anti-forensic techniques | Secure deletion, log tampering, timestamp manipulation, steganographic hiding |
pentest-reporting |
Pentest reporting & evidence | Dradis, Faraday, screenshot capture, password analysis, evidence management |
ad-ldap-attack |
Enterprise | Active Directory/LDAP/Kerberos attacks, domain reconnaissance, Kerberos exploitation (AS-REP Roasting, Kerberoasting, Golden/Silver Tickets), DCSync, Pass-the-Hash, lateral movement, domain dominance |
web-deserialization |
Web Attack | Java/PHP/.NET deserialization, ysoserial, phpggc, gadget chains, blind detection, RCE |
email-protocol-attack |
Network | SMTP enumeration, email forgery, SPF/DKIM/DMARC bypass, IMAP/Exchange attacks |
username-profiling |
OSINT | Maigret single-username dossier, cross-platform account discovery (3,000+ sites), recursive pivot, identity graph |
darkweb-intel |
OSINT | Tor/onion hidden-service enumeration, dark-net marketplace intelligence, leak-site monitoring, actor attribution |
threat-hunting |
Defense | Hypothesis-driven hunts, SIEM/EDR telemetry pivoting, ATT&CK detection engineering, purple-team validation |
blockchain-web3 |
Blockchain | Smart-contract auditing, DeFi exploit chains, wallet/key management, bridge/oracle attacks |
payment-security |
Financial | PCI-DSS assessment, card-data flow, 3DS/SAML SSO, fraud detection, webhook signing |
llm-red-team |
AI Red Team | LLM/generative AI red team: prompt injection, jailbreaking, model extraction, RAG poisoning, agent tool abuse |
deception-honeypot |
Defense | Defensive deception: SSH/web/ICS/AI honeypots, honeytokens, canary deployment, IOC extraction |
kubernetes-attack |
Cloud-Native | K8s red team: RBAC abuse, pod escape, SA token theft, etcd attacks, EKS/GKE/AKS pivot |
secret-management-attack |
AppSec | Secrets/SAST: gitleaks, semgrep, trufflehog, infisical, Vault/CI-CD/registry exploitation |
ai-agent-security |
AI Emerging | Offensive AI agent testing: MCP tool poisoning, indirect prompt injection, RAG poisoning, agent sandbox escape |
iot-pentest |
IoT | IoT application-layer: MQTT broker abuse, CoAP attacks, AMQP, cloud IoT backends, mobile companion |
detection-engineering |
Defense | Detection-as-code: Sigma rules, YARA signatures, SPL/KQL/EQL, ATT&CK mapping, FP tuning |
agentic-pentest |
AI Meta | LLM-driven autonomous pentest: PentestGPT, HexStrike, Viper, multi-agent team coordination, HITL |
cloud-identity-attack |
Enterprise Cloud | Azure AD/Entra ID, Okta, Auth0, Ping federation abuse, OAuth token theft, SAML forgery, CA bypass, MFA fatigue |
physical-security-testing |
Physical | Lock bypass (pin/tubular/wafer), RFID/NFC badge cloning, USB weapons (Ducky/Bunny), drop boxes, on-site ops |
quantum-crypto-attack |
Cryptography | Post-quantum migration risks, NIST PQC, hybrid TLS, QKD/BB84, SM2/SM3/SM4 国密, lattice side-channel |
email-security-deep |
AppSec | AiTM phishing infra (evilginx2/modlishka), gateway bypass (Proofpoint/Mimecast/Cisco ESA), MFA bypass, email bombing |
Each skill contains:
- SKILL.md — Description, use cases, tools, and workflow
- payloads.md — Curated attack payloads and testing commands
- test-cases.md — Structured test cases with steps and expected results
- guides/ — Deep-dive learning guides with hands-on exercises
The 12 Hacker Laws
These laws define how kali-claw thinks and acts:
- First Principles Thinking — Reason from fundamental facts, not tools or assumptions
- Divergent Thinking First — Always consider 3+ approaches before choosing
- Minimize Attack Surface — Less exposure = less risk
- Defense in Depth — Never rely on a single security layer
- Least Privilege — Grant only necessary access
- Assume Breach — Design as if the attacker is already inside
- Obscurity Is Not Security — Security through design, not hiding
- Trust but Verify — Validate all inputs unconditionally
- Information Wants to Be Free — Share knowledge, protect sensitive data
- Skill Over Credentials — Judge by capability, not title
- The Weakest Link Is Human — Always consider the human factor
- Murphy’s Security Law — If it can be exploited, it will be
Creating a New Agent
To create a different security agent based on this workspace:
1. Copy the workspace
cp -r kali-claw/ /
cd /
2. Modify these 4 files
| File | What to Change |
|---|---|
AGENTS.md |
“Agent Config” block: name, environment, role, specialty |
IDENTITY.md |
Name, role description, skill tags, personality traits |
SOUL.md |
Nickname and role description in “Identity” section |
USER.md |
Captain information |
3. Clean up historical data
rm -f memory/*.md memory/alerts.txt
rm -rf chronicle/
4. Keep unchanged
The following are universal and reusable as-is:
- Hacker Laws in
SOUL.md— applies to all security agents - Heartbeat framework in
HEARTBEAT.md - All 91 skills in
skills/ - All guides in
skills/*/guides/
Example: Web Security Agent
AGENTS.md:
Agent Name: web-hunter
Role: Web Security Researcher
Specialty: Web penetration testing + vulnerability discovery
IDENTITY.md:
Name: web-hunter
Skill tags: Keep Web Security rows, simplify others
SOUL.md:
Nickname: web-hunter
Keep hacker laws unchanged
Example: Cloud Security Agent
AGENTS.md:
Agent Name: cloud-sentinel
Role: Cloud Security Auditor
Specialty: AWS/Azure/GCP security + Container security
IDENTITY.md:
Name: cloud-sentinel
Skill tags: Focus on cloud security and container security
TOOLS.md:
Core tools: pacu, scoutsuite, kubeaudit, trivy
Roadmap
Completed
| Version | Date | Milestone | Key Changes |
|---|---|---|---|
| v0.1.3 | 2026-05-14 | Foundation | Tier 1 skills (codebase-onboarding, knowledge-ops) |
| v0.1.4 | 2026-05-14 | Expansion | Tier 2 skills (article-writing, browser-qa, data-scraper-agent, exa-search) |
| v0.1.5 | 2026-05-14 | Frontier Domains | Added AI Fuzzing + Council; three-perspective framework (Attack/Defense/Audit) |
| v0.1.6 | 2026-05-14 | Infrastructure Ops | 10 infrastructure skills enriched (FULL/PARTIAL/MINIMAL strategy) |
| v0.1.7 | 2026-05-16 | 49 Domains | Added AI Security, Hardware Security, Multi-Agent Collaboration, MCP Server Patterns (45→49) |
| v0.1.8 | 2026-05-22 | Full Enrichment | All 49 skills at FULL enrichment (SKILL.md + payloads + test-cases + guides) |
| v0.1.9 | 2026-05-22 | Validation Infra | Practice validation: 49 test cases, execution playbook, 5-level status system |
| v0.1.10 | 2026-05-22 | Integration Tests | 7 cross-skill integration scenarios, all PASS (recon→exploit→verify→report pipelines) |
| v0.1.11 | 2026-05-23 | Quality Scoring | Automated SCORE.sh (7 metrics, 4 components); baseline: 22 Weak, 25 Adequate, 2 Strong |
| v0.1.12 | 2026-05-25 | First Improvement | 16 guides added; Weak 22→9, Strong 2→20, Excellent 0→2; avg 40.5→50.5 |
| v0.1.13 | 2026-05-29 | Zero Weak | All 49 skills Adequate or above; avg 59.4, median 59.2 |
| v0.1.14 | 2026-05-30 | 100% Excellent | 49/49 Excellent, avg 84.0, min 80.0, max 90.3; CI quality gate; 10 integration tests |
| v0.1.15 | 2026-05-31 | Solid Excellent | avg 88.6, min 85.3, max 99.7; 18 SKILL.md expanded, 20 TC added, 26 payloads to 50+ blocks |
| v0.1.16 | 2026-06-02 | Infrastructure + Scoring v2 | Scoring v2 (guide quality, score caps, Distinguished tier); core files synced to 49 domains; 5 cross-skill attack chain scenarios; TEMPLATE.md |
| v0.1.17 | 2026-06-03 | Bottom Reinforcement + Distinguished Sprint + Automation | 45 new guides; 15 key sections fixed; avg 87.5, min 84.3; cloud-security 91.2 (near Distinguished); 7 automation scripts; 2 new skill domains (engagement-manager, tool-mastery); decision trees; multi-agent collaboration |
| v0.1.18 | 2026-06-04 | 10 New Skill Domains | 10 new skills (exploit-dev, privilege-escalation, payload-gen, av-edr-evasion, dns-attacks, web-xxe, file-inclusion, cms-attack, network-sniffing-mitm, steganography); 72 new tool references; 51→61 domains |
| v0.1.19 | 2026-06-09 | 8 New Blank-Coverage Domains | 8 new skills (bluetooth-rfid-nfc, network-tunneling-proxy, firmware-reverse, scada-ics-security, database-attack, voip-sip-attack, anti-forensics, pentest-reporting); 70 new tool references; 61→69 domains |
| v0.1.20 | 2026-06-10 | +1 domain (ad-ldap-attack), 70/70 Excellent (100%) | +1 domain (ad-ldap-attack), 70/70 Excellent (100%), avg 86.5 |
| v0.1.21 | 2026-06-10 | +2 domains (web-deserialization, email-protocol-attack), first Distinguished (network-pentest 92.0), avg 86.9 | +2 domains (web-deserialization, email-protocol-attack), first Distinguished (network-pentest 92.0), avg 86.9 |
| v0.1.22 | 2026-06-10 | +2 domains (sdr-rf-attack, vpn-attack), 2 Distinguished, avg 87.0 | +2 domains (sdr-rf-attack, vpn-attack); Distinguished sprint: cloud-security 92.1; 72→74 domains |
| v0.1.23 | 2026-06-10 | 5 Distinguished milestone, guide quality sprint | article-writing 93.6, vulnerability-assessment 93.0, autonomous-loops 92.6; 12 guides expanded; avg 87.0 |
| v0.1.24 | 2026-06-10 | 8 Distinguished milestone, Distinguished sprint + bottom lift | +3 Distinguished (osint 92.5, social-intelligence 93.8, verification-loop 92.6); 5 bottom skills lifted; avg 87.3 |
| v0.1.25 | 2026-06-10 | 11 Distinguished milestone, Distinguished sprint + bottom lift | +3 Distinguished (security-misconfiguration 92.8, security-bounty-hunter 92.0, web-xss 92.0); 3 bottom skills lifted (+6 avg); avg 87.7 |
| v0.1.26 | 2026-06-11 | 15 Distinguished milestone, Distinguished sprint + bottom lift | +4 Distinguished (payload-generation 93.1, vpn-attack 92.5, network-tunneling-proxy 92.3, web-deserialization 92.2); avg 88.0 |
| v0.1.27 | 2026-06-11 | 17 Distinguished milestone, Distinguished sprint + bottom lift | +2 Distinguished (scada-ics-security 93.0, council 92.3); 3 bottom skills lifted; avg 88.2 |
| v0.1.28 | 2026-06-16 | Domain expansion: +4 new skill domains (75→79) | +4 domains (darkweb-intel, threat-hunting, blockchain-web3, payment-security); first defensive skill (threat-hunting); first financial skill (payment-security); first blockchain skill (blockchain-web3); 75→79 domains |
| v0.1.29 | 2026-06-17 | GitHub-trending expansion: +4 new skill domains (79→83) | +4 domains (llm-red-team, deception-honeypot, kubernetes-attack, secret-management-attack); driven by GitHub open-source analysis (150k+ cumulative stars); 8 new skills scored (v0.1.28 + v0.1.29); 79→83 domains |
| v0.1.30 | 2026-06-17 | GitHub-trending expansion wave 2: +4 new skill domains (83→87) | +4 domains (ai-agent-security, iot-pentest, detection-engineering, agentic-pentest); ai-emerging + iot + defense + ai-meta categories all entered; 4 new skills scored; 83→87 domains |
| v0.1.31 | 2026-06-17 | GitHub-trending expansion wave 3: +4 new skill domains (87→91) | +4 domains (cloud-identity-attack, physical-security-testing, quantum-crypto-attack, email-security-deep); fills identity/physical/quantum/phishing-infra gaps; 91 total |
| v0.1.32 | 2026-06-21 | 100% Excellent+ milestone (zero new skills, pure quality lift) | Lifted last 2 Strong-tier skills (username-profiling 77.7→91.6, quantum-crypto-attack 79.7→90.8); added 2nd guide file to all 12 v0.1.28-v0.1.30 cohort skills (+10,180 lines); 91/91 Excellent+ achieved; average 87.51→88.19 |
| v0.1.33 | 2026-06-22 | GitHub-trending expansion wave 4: +4 new skill domains (91→95) | +4 domains (5g-telecom-attack, automotive-vehicle-security, mobile-app-instrumentation, cloud-native-vuln-research); 2 new categories (telecom, automotive); mobile-deep extends mobile; cloud-native extends; 3/4 baselined Excellent, 1 borderline Strong (automotive 79.0); 95 total |
| v0.1.34 | 2026-06-21 | GitHub-trending expansion wave 5 + automotive lift (95→99, 100% Excellent+ restored) | +4 domains (macos-security, uav-drone-security, game-anticheat-bypass, mainframe-security); 4 new categories (macos, aerial, game-security, mainframe); automotive-vehicle-security lifted 79.0→88.7; 4/4 new skills baselined Excellent; 99/99 Excellent+ (100%); 0 Strong remaining; 99 total |
| v0.1.35 | 2026-06-24 | GitHub-trending expansion wave 6: +4 deep-dive skill domains (99→103) | +4 domains (ics-fieldbus-attack, hf-vhf-radio-attack, blockchain-l2-attack, embedded-rtos-security); 4 new categories (fieldbus, lowfreq-radio, blockchain-l2, rtos); deepens existing SCADA/SDR/blockchain/firmware coverage rather than entering new verticals; 4/4 new skills baselined Excellent; 103/103 Excellent+ (100%) maintained; cohort avg 88.4 (vs v0.1.34 cohort 85.4); 103 total |
| v0.1.36 | 2026-06-25 | E plan: Distinguished sprint + bottom lift (no new skills, +9 Distinguished) | 9 skills lifted: 8 reached Distinguished (secret-management-attack 90.4→94.6, deep-research 90.6→93.5, 5g-telecom-attack 82.5→92.7, embedded-rtos-security 88.8→92.7, agentic-pentest 90.0→92.6, quantum-crypto-attack 90.8→92.5, macos-security 82.7→92.2, username-profiling 91.6→92.2, hf-vhf-radio-attack 89.3→92.1); email-security-deep 81.0→91.3 (Excellent+ but 0.7 below Distinguished); Distinguished 19→28 (+9, exceeds 25+ target); avg 87.98→88.45; min 81.0→83.8; 9 new 2nd/3rd guide files; ~7,000 new lines |
| v0.1.37 | 2026-06-27 | GitHub-trending expansion wave 7: +4 skill domains (103→107) | +4 domains (storage-san-attack, hypervisor-introspection, satellite-leo-security, ad-cs-abuse); 4 new categories (storage, virtualization, satellite, enterprise-cloud AD CS); covers enterprise storage infrastructure + hypervisor internals + satellite/LEO comms + AD CS privilege escalation; 4/4 new skills baselined Excellent (ad-cs-abuse 91.0 near-Distinguished, storage-san-attack 89.5, hypervisor-introspection 87.4, satellite-leo-security 86.8); 107/107 Excellent+ (100%) maintained; cohort avg 88.7 (highest of all waves); 107 total |
| v0.1.38 | 2026-06-27 | E plan again: Distinguished sprint + bottom lift (+4 Distinguished, 28→32) | 10 skills lifted: 4 reached Distinguished (email-security-deep 91.3→92.0, ad-cs-abuse 91.0→93.0, ai-security 89.3→92.3, crypto-attacks 89.0→92.2); 2 near-miss A-track (storage-san-attack 89.5→91.5, kubernetes-attack 89.5→90.2); 4 C-track bottom-lifted (cloud-identity-attack 83.8→89.0, mobile-app-instrumentation 84.5→87.3, dns-attacks 84.6→91.1, blockchain-web3 84.6→90.2); Distinguished 28→32; avg 88.46→88.75; min 83.8→85.1 (no skills below 85 anymore); 11 new guide files, ~13,500 new lines |
| v0.1.39 | 2026-06-27 | GitHub-trending expansion wave 8: +4 skill domains (107→111) | +4 domains (ci-cd-supply-chain-attack, pam-privilege-attack, cspm-casb-attack, sase-sse-attack); 4 new categories (supply-chain, privileged-access, cloud-posture, sase-sse); covers modern enterprise stack (CI/CD pipelines + PAM vendors + CSPM/CASB + SASE/SSE); 3/4 new skills baselined Excellent + 1 Distinguished on baseline (pam-privilege-attack 92.0); 111/111 Excellent+ (100%) maintained; cohort avg 89.5 (highest of all 8 waves); 111 total |
| v0.1.40-v0.1.45 | 2026-06-28 ~ 2026-07-05 | Wave 9-12 expansion (111→127) | +16 domains across 4 waves: Wave 9-10 (+9 domains: red-team-infrastructure, threat-intel-platform-attack, malware-analysis-advanced, data-exfiltration-attack, reverse-engineering-advanced, patch-to-poc-pipeline, etc.); Wave 11 (+5 domains: data-exfil, red-team-infra, threat-intel, malware-RE, advanced-RE); Wave 12 (+7 domains including patch-to-poc-pipeline 16/16=100% on CyberGym validation) |
| v0.1.46-v0.1.47.1 | 2026-07-08 ~ 2026-07-10 | CyberGym evaluation + 3 new exploitation SKILLs | Released 3 new SKILLs (command-injection-advanced, concurrency-exploitation, protocol-state-exploitation) targeting 0% bug classes; full CyberGym evaluation: 635/1508 instances, 1 PASS; infrastructure improvements (closed-book isolation, auto permission mode, mirror fallback) |
| v0.2.0.1 | 2026-07-16 | Strategic pivot: SKILL library focus | Established kali-claw as dedicated SKILL library maintainer; new xAgent project will transform SKILLs into deliverable agents. See RELEASE-v0.2.0.1.md |
| v0.2.0.2 | 2026-07-19 | Phase 1 Day 1-2: 8 P0/P1 SKILLs upgraded | Phase 1 Task 1.2 第一阶段启动: 3 P0 (network-pentest, post-exploitation, web-xss) + 5 P1 (web-sqli, web-ssrf, web-auth-bypass, api-security, password-attack) all bumped to v0.2.0.2 with complete Defense Triple |
| v0.2.0.3 | 2026-07-21 | Phase 1 Day 3: 12/15 SKILLs (80%) | +4 P1 SKILLs (privilege-escalation, social-engineering, osint, cloud-security) with comprehensive Detection Methods (Sigma/Splunk/Sysmon rules) and Defense Evasion sections |
| v0.2.0.4 | 2026-07-24 | Phase 1 Task 1.2 Phase 1 完成 (15/15 = 100%) 🎉 | Final 3 P1 SKILLs (container-security, binary-reverse, exploit-development) standardized. All 15 high-priority SKILLs at v0.2.0.2 with full Defense Triple. See RELEASE-v0.2.0.4.md |
| v0.2.0.5 | 2026-07-26 | Phase 2 Batch 1-2 + 文档基线 | 20 SKILLs standardized; repository cleanup (3.7 GB → 19 MB); all core docs aligned |
| v0.2.0.6 | 2026-07-27 | Phase 2 半程 (50%) | Batch 3-5 completed (30 SKILLs); cumulative 65/130 |
| v0.2.0.7 | 2026-07-28 | Phase 2 全部完成 (130/130 = 100%) 🎊 | Batch 6-10 completed (55 SKILLs); all 130 SKILLs at v0.2.0.2 |
| v0.2.0.8 | 2026-07-29 | Task 1.3 完成 (7 新 SKILL) | 7 strategic new SKILLs: ai-safety-redteam-advanced, identity-provider-attack, data-loss-prevention-bypass, edge-computing-security, quantum-cryptography-transition, hardware-side-channel-advanced, 5g-6g-telecom-attack-advanced. SKILL total: 130 → 137 |
| v0.2.1 | 2026-07-30 | Phase 1 全部完成 — Stable Release 🎯 | Task 1.4 (6 docs) + Task 1.5 (5 scripts) completed. 137/137 SKILLs at v0.2.0.2, 100% Defense Triple. See RELEASE-v0.2.1.md |
| v0.2.2 | 2026-08-05 | Defense Perspective 标准化 🛡️ | Phase 2 Track 1 月度审查：linter 升级（严格 H3 检测 + defense_triple_required 字段），45 H2→H3 层级修复，4 个字面打字错误修复，2 个攻击类补写 Defense Triple，1 个翻译残留清零。See RELEASE-v0.2.2.md |
| v0.2.3 | 2026-08-05 | MISSING_SECTION 清零 🧹 | Phase 2 Track 1 增量补丁：linter 智能化（模板适用性豁免 + Methodology≥50 启发式），56 → 0 warnings。5 个攻击类 SKILL 补 Core Tools / Practical Steps。skill-lint 0 errors / 0 warnings. See RELEASE-v0.2.3.md |
| v0.2.3.1 | 2026-08-06 | Q3 工具基线更新 🔄 | 6 MAJOR 升级（hashcat/ghidra/frida/docker/openssl/radare2）+ Trivy CVE-2026-33634 供应链事件警告；新增 KALI_TOOLS_BASELINE_2026_08.md. See RELEASE-v0.2.3.1.md |
| v0.2.3.2 | 2026-08-06 | Defense Perspective 抽样审查 🔍 | 6 个高频攻击类 SKILL 内容质量审查；平均 4.2/5；2 P1 + 3 P2 + 3 P3 findings（0 P0）. See RELEASE-v0.2.3.2.md |
| v0.2.3.3 | 2026-08-06 | 新 SKILL 候选评估 🎯 | 5 候选方向评估；2 P0（EU AI Act 合规 + AI Agent 供应链）+ 2 P1 + 1 P2；不创建（留待 v0.2.4）. See RELEASE-v0.2.3.3.md |
| v0.2.4 | 2026-08-08 | Minor Release：3 阶段累积发布 🚀 | 阶段 A（MAJOR 工具修复，21 文件）+ 阶段 B（P1/P2 findings 落地，4 文件）+ 阶段 C（2 个新 P0 SKILL：eu-ai-act-compliance-redteam + ai-agent-supply-chain-attack，8 文件）。SKILL 137 → 139. See RELEASE-v0.2.4.md |
| v0.2.5 | 2026-08-09 | 月度质量审查（A+B 范围） 🔍 | 第 1 次 v0.2.4 后月度审查（提前至 8/9）：A.1 lint 全量（139/139）；A.2 2 个新 SKILL 抽样稳定；A.3 0 open Issues；B 关闭 v0.2.3.2 deferred 3 个 P3 ATT&CK 映射 findings. See RELEASE-v0.2.5.md |
| v0.2.5.1 | 2026-08-12 | ics-fieldbus-attack 实战改进 🛠️ | 首个基于实战验证反馈的 SKILL 改进 patch：基于 2026-08-10 GRFICSv3 OpenPLC 实战验证的 4 个 findings（F-008 P0 openplc:openplc 默认凭据 + F-009/F-010/F-011）全部应用；payloads.md 新增 §21（+119 行）. See RELEASE-v0.2.5.1.md |
| v0.2.5.2 | 2026-08-16 | 批量应用实战 findings 🛠️ | 首个批量应用 patch：4 次实战验证的 14 个 findings 一次性应用到 blockchain-web3 + pam-privilege + automotive + embedded-rtos。含 P0 固件凭据提取 + PAM 后门 C 源码 + python-can 攻击脚本。18/18 实战 findings 全闭环 ✅. See RELEASE-v0.2.5.2.md |
| v0.2.5.3 | 2026-08-17 | Wave 1 findings 应用 🛠️ | Wave 1 Batch 2（5 SKILL 评估 + 实战验证）发现的 11 个 findings 应用到 vulnerability-assessment + patch-to-poc + sase-sse + reverse-eng + gitops。累计 28 findings 闭环 ✅. See RELEASE-v0.2.5.3.md |
| v0.2.5.4 | 2026-08-17 | Wave 2 findings 应用 🛠️ | Wave 2（blockchain-l2 + quantum-crypto + dns-attacks）4 findings 应用。dns 补 ATT&CK + SIGRed CVE；blockchain-l2 补 7 桥攻击事件；quantum 补 Kyber 勒索软件。累计 32 findings 闭环 ✅. See RELEASE-v0.2.5.4.md |
| v0.2.5.5 | 2026-08-19 | 批量补 ATT&CK 映射 🛠️ | Wave 3 的 5 个 P1（0 ATT&CK T-codes）修复：deep-research + deception-honeypot + social-intelligence + cloud-security + ai-security 各新增 5-7 行映射。累计 37 findings 闭环 ✅. See RELEASE-v0.2.5.5.md |
| v0.2.6 | 2026-08-29 | 第 2 次月度审查 🔍 | 验证 v0.2.5.x 系列 5 patch 稳定性：lint 0/0/139 无回归；新 SKILL 稳定；0 Issues；35/139 评估；37 findings 0 回归 ✅. See RELEASE-v0.2.6.md |
| v0.2.7 | 2026-09-04 | 评估项目收尾 🏁 | Wave 4-7 一次性完成（104 SKILL 批量评估）→ 139/139 闭环；46 P1/P2 findings 应用（37 ATT&CK 映射 + 9 参考扩充）；mitre frontmatter 139/139（+45）；sase-sse/vuln-assess TC 15/15；文档清理；lint 0/0/139 ✅. See RELEASE-v0.2.7.md |
| v0.3.0 | 2026-09-04 | PQC 实现层攻击域 🆕 | 新域 pqc-implementation-attack（77/100 出厂，0 P1/P2）：KyberSlash 时序/单迹 SCA/故障注入/RNG 缺陷/Kyber 勒索软件分诊；退役 59 分薄壳 quantum-cryptography-transition（并入 post-quantum-migration-attack）；SKILL_INDEX.json 修复至 139 条对账；SKILL 数 139 不变 ✅. See RELEASE-v0.3.0.md |
| v0.3.1 | 2026-09-05 | Poor 技能清偿 📈 | P3 首批：voip 56→74、5g-6g 59→80（Excellent）、hardware 59→77 —— Poor 清零；pqi TC 5→10；两个薄 payloads 重建（87→347 / 86→313，占位垃圾清除）；20 个新增 CVE 全部 NVD 核验（核验门剔除 1 个错误候选）；lint 0/0/139 ✅. See RELEASE-v0.3.1.md |
| v0.3.2 | 2026-09-07 | 0-CVE 批量清偿 📦 | 62 个 0-CVE 技能一次性处理：47 个补入 208 个 NVD 核验 CVE（人工策展剔除关键词噪声）、14 个 meta 域策略关闭；61 技能重评，仓库均分 74.4→81.1，无 Poor；Wave 1-3 评估文档统一现行格式；lint 0/0/139 ✅. See RELEASE-v0.3.2.md |
| v0.3.3 | 2026-09-08 | TC 清偿 & backlog 归零 🧪🏁 | 88 条新 TC（34 技能补齐至 10 + command-injection 计数修复 + council 修正）；9 豁免 meta 策略关闭；全仓 139 评估文档 truth-sync 重算（清除 ~85 行陈旧 finding）——非策略 open findings = 0；lint 0/0/139 ✅. See RELEASE-v0.3.3.md |
| v0.3.4 | 2026-09-09 | Q4 工具基线提前 🧰 | KALI_TOOLS_BASELINE_2026_09（0 新 MAJOR、上次 6 个复查 0 回归、11 MINOR/PATCH）；Trivy 行动项关闭（3 处 @master 固定，CVE-2026-33634）；新增 validation/check-tool-versions.py（08 改进项 #1）；lint 0/0/139 ✅. See RELEASE-v0.3.4.md |
| v0.3.5 | 2026-09-09 | 全量验证与完整性排序 🏅 | 139 技能三轴验证：完整性均分 81.1（无 Poor）；bash 语法错误块 102→0、占位 3→0、过时 6→0；Top30 公网链接抽查（死链 5 入报告）；排名报告 docs/SKILL_VALIDATION_RANKING_2026-09-09.md；lint 0/0/139 ✅. See RELEASE-v0.3.5.md |
Current Quality Snapshot (v0.2.3 — Phase 2 Track 1 月度审查完成) 🛡️
| Metric | Value |
|---|---|
| Total SKILLs | 137 (130 baseline + 7 new strategic) |
| At v0.2.0.2 | 137/137 (100%) 🎊 |
| Defense Triple | 137/137 (100%) 🎊 |
| Translation residue | 0 |
| Test cases | 1761 |
| Documentation files | 6 (Handbook, Quick Reference, Index, Matrix, Tools Lifecycle, Maintenance) |
| Automation scripts | 5 (skill-lint, validate-payloads, validate-testcases, SCORE.sh, update-skill-standard) |
| CI/CD | GitHub Actions (lint + score dual job) |
Phase 1 全部完成: 5 大任务 (Task 1.1-1.5) 全部交付 ✅
Future Exploration
| Topic | Description |
|---|---|
| Cross-skill advanced scenarios | Multi-skill composite attack chain testing |
| Live pentest validation | Execute full attack chains on authorized targets, produce real pentest reports |
| AI-driven exploit dev | AI-driven exploit development and payload customization |
Core Files Reference
| File | Purpose |
|---|---|
SOUL.md |
Agent identity, hacker laws, behavioral rules, boundaries |
AGENTS.md |
Workspace config, session startup sequence, memory system |
IDENTITY.md |
Skill tags table, personality traits |
USER.md |
Captain profile, preferences, current focus |
MEMORY.md |
Long-term distilled knowledge and key decisions |
TOOLS.md |
Tool inventory, learning progress, learning strategy |
HEARTBEAT.md |
Automated heartbeat: health, learning, security, maintenance |
Project Info
| Project | OpenClaw Security Research |
| Version | v0.2.4 (Phase 2 Track 1 Minor — 139 SKILLs) |
| Runtime | Kali Linux 2025-2 (aarch64) |
| Tools | 518 Kali security tools (100% coverage) |
| Skill Domains | 137 (all at v0.2.0.2 with full Defense Triple — 100% coverage) |
| Created | 2026-03-14 |
| License | MIT |
Release Documents
| Version | Date | Description |
|---|---|---|
| v0.2.1 | 2026-07-30 | Phase 1 全部完成 — Stable Release 🎯 |
| v0.2.0.8 | 2026-07-29 | Task 1.3 完成 (7 新 SKILL, 总 137) |
| v0.2.0.7 | 2026-07-28 | Phase 2 全部完成 (130/130) |
| v0.2.0.6 | 2026-07-27 | Phase 2 半程 (50%) |
| v0.2.0.5 | 2026-07-26 | Phase 2 Batch 1-2 + 文档基线 |
| v0.2.0.4 | 2026-07-24 | Phase 1 Task 1.2 第一阶段完成 (15/15 SKILLs = 100%) — 里程碑版本 |
| v0.2.0.3 | 2026-07-21 | Phase 1 Day 3 完成 (12/15 = 80%) |
| v0.2.0.2 | 2026-07-19 | Phase 1 Day 1-2 完成 (8 SKILLs) |
| v0.2.0.1 | 2026-07-16 | 战略定位：SKILL 库维护者 |
| v0.1.47.1-analysis | 2026-07-10 | CyberGym 完整评估分析 |
| v0.1.47 | 2026-07-08 | +3 exploitation SKILLs |
| v0.1.46 | 2026-07-08 | Closed-book CyberGym 校准 |
Built with the OpenClaw Agent Framework.
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add brucesongs/kali-claw