Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Security
Guidance for securing Infrastructure-as-Code (Bicep, ARM, Terraform, and ACI/Container) pipelines on Azure — shift-left scanning, policy-as-code, secret hygiene, identity for pipelines, drift detection, and supply chain. Covers Microsoft Defender for Cloud DevOps Security (GitHub / Azure DevOps connectors), Microsoft Security DevOps (MSDO) extension, Bicep linter and template specs, Terraform best practices on Azure (state management, backend hardening, provider versions), Azure Policy as deploy-time and CI-time gate, GitOps with Flux/ArgoCD on AKS, secret scanning and dependency review, signed commits / artifact signing, OIDC federation for pipeline identity (no client secrets), and integration with Defender for Cloud posture. WHEN: IaC security, Bicep security, Terraform Azure security, Defender for Cloud DevOps Security, MSDO, Microsoft Security DevOps, shift-left Azure, policy as code, OIDC GitHub Actions Azure, secrets in IaC, drift detection, signed Bicep, supply chain Azure. DO NOT USE for Kubernetes admission control alone (use defender-for-containers), generic CI/CD design, or app-layer security.
Implements envelope encryption with AWS KMS, encrypting data locally with a data encryption key (DEK) and protecting that DEK with a KMS-managed key (KEK), covering the encrypt/decrypt flow, KMS key types, and security validation criteria. Use when designing key management for encrypting large or numerous data objects on AWS, or when reducing direct KMS API call volume.
Integrates AFL++ coverage-guided fuzzing into CI/CD pipelines, covering harness construction, AFL++/AddressSanitizer/CmpLog instrumentation builds, and persistent-mode fuzzing to discover memory-corruption and input-handling vulnerabilities in C/C++ code. Use when adding automated fuzz testing to a build pipeline or hunting for memory-safety bugs in native/compiled applications.
IngressNightmare (CVE-2025-1974) — unauthenticated RCE in the Kubernetes Ingress-NGINX Admission Controller. Use when the target K8s cluster uses ingress-nginx, an admission webhook is found on ports 443/8443, or the admission controller is reachable over the Pod network. Covers the vulnerability's mechanics, exploitation precondition assessment, PoC usage, and follow-up lateral movement.
Network reconnaissance and service discovery within Kubernetes clusters. Use after obtaining a Pod shell when discovering other services in the cluster or scanning the internal K8s network. Covers DNS PTR reverse lookups, SRV record enumeration, AXFR zone transfers, and using K8Spider. Use this skill for lateral reconnaissance in a Pod, finding hidden services, or identifying targets, even if the user does not explicitly mention DNS.
Local pentest sandbox for a full black-box engagement. Triggers on "kage", "pentest", "security audit on", "audit the security of". Runs recon, deep testing, exploit verification, and judging inside a per-engagement Kali Docker container. Each host working directory gets its own isolated sandbox. Produces `./results/<target>/audit-report.md`.
Use masscan for very large-scale, high-speed port scanning. Use when scanning large CIDR ranges or at whole-internet scale. masscan is an asynchronous, stateless scanner capable of millions of packets per second, suited to large-scale asset discovery. Requires root privileges. Use this skill for any scenario involving large-scale network scanning, high-speed port discovery, or internet-wide mapping
A methodology for network-layer poisoning and protocol spoofing attacks during internal network penetration testing. Use when network poisoning is needed to obtain NTLM authentication or carry out a man-in-the-middle attack. Covers ARP poisoning, DHCP poisoning, DHCPv6/mitm6 spoofing, DNS spoofing, ADIDNS record injection, WPAD hijacking, WSUS poisoning, ICMP redirection, and NBT Name Overwrite. Usually used as a prerequisite for an NTLM relay attack.
Use NetExec (nxc) / CrackMapExec (cme) for bulk Windows/AD network operations. Use when you need bulk credential validation, password spraying, remote command execution, share enumeration, or credential dumping. nxc is the successor to CrackMapExec and supports SMB/WinRM/LDAP/MSSQL/SSH/RDP/FTP/WMI and other protocols. In domain penetration testing it is used for bulk operations — validating credentials or executing commands across an entire subnet with a single command. Use this skill for scenarios involving CrackMapExec, NetExec, nxc, cme, password spraying, bulk SMB, or domain user enumeration
Comprehensive OSINT methodology for external red-team operations and authorized attack-surface assessments. Covers the 5-stage recon pipeline (seed discovery, asset expansion, enrichment, exposure analysis, reporting), asset-graph discipline with 29 asset types, severity rubric (CRITICAL/HIGH/MEDIUM/LOW/INFO), confidence upgrade workflows, time budgeting, asset-level triage rules, scale-based tactics, identity-fabric mapping (Entra/Okta/ADFS/Google/SAML/M365 Teams+SharePoint+OAuth), API and auth-map methodology, JavaScript deep analysis, mobile attack surface, cloud attack surface, breach×identity correlation, detectability tagging, detection-aware probing (back-off, persona rotation), read-only validator discipline, WAF/CDN bypass + origin discovery, vulnerability prioritization (CVE/EPSS/KEV), phishing infrastructure planning + pretext development, bug bounty submission templates, client deliverable templates with risk translation, threat-actor investigation (incl. RU/CN pivots), cryptocurrency tracing, image/video forensics, chronolocation. Use when planning or executing reconnaissance against authorized targets, mapping an organization's external attack surface, investigating a person/entity, tracing crypto flows, geolocating media, or performing attribution work.
Use PadBuster to perform Padding Oracle attacks. Use when a web application uses CBC-mode encryption and has a Padding Oracle vulnerability. PadBuster can automatically decrypt ciphertext and forge valid ciphertext for arbitrary plaintext, making it suitable for encrypted cookies, tokens, and URL parameters. Use this skill for any scenario involving Padding Oracle attacks, CBC ciphertext decryption, or cookie forgery
Guidance for Microsoft Purview AI Hub (now part of Data Security Posture Management for AI) — discover, govern, and protect sensitive data flowing into AI applications (Microsoft 365 Copilot, Copilot Studio agents, ChatGPT, Gemini, third-party generative AI). Covers AI app discovery via Defender for Cloud Apps + endpoint signals, sensitive data risk surface, ready-to-use policies for Copilot oversharing and risky AI usage, DLP for generative AI endpoints (browser blocking), prompt/response auditing, integration with IRM (risky AI usage), Sentinel reporting, and difference vs Defender for Cloud AI workload protection. WHEN: Purview AI Hub, DSPM for AI, AI app discovery, ChatGPT data leakage, Gemini DLP, generative AI risk, prompt audit, Copilot Studio agent governance, sensitive data to AI, shadow AI, agent data risk. DO NOT USE for Defender for Cloud's AI workload protection (use defender-for-cloud-ai), Azure AI Content Safety (use azure-ai-content-safety), or Microsoft 365 Copilot rollout (use copilot-for-m365-readiness).
Guidance for Microsoft Purview Customer Key and Double Key Encryption (DKE) — customer-controlled encryption for service-side data (Customer Key for Exchange/SharePoint/OneDrive/Teams/Purview) and client-side double encryption for highly sensitive documents (DKE via sensitivity labels and a customer-hosted key service). Covers Azure Key Vault setup (HSM-backed, recoverable, soft-delete, purge protection, redundant regions), data encryption policies (DEPs) for M365 services, key rotation and revocation, DKE service deployment (containerized key release service), DKE label authoring, end-user experience trade-offs (no eDiscovery into DKE content), and decision criteria vs Microsoft-managed encryption. WHEN: Customer Key M365, BYOK Exchange, DKE Microsoft, double key encryption, customer-managed keys SharePoint OneDrive Teams, key rotation Customer Key, HYOK alternative, sovereignty encryption, regulator-controlled encryption, sensitivity label DKE. DO NOT USE for Azure resource CMK (use azure-key-vault), general sensitivity labels (use purview-information-governance), or non-encryption compliance (use compliance-manager).
Guidance for Microsoft Purview Insider Risk Management (IRM) — detect, investigate, and act on risky user activity (data theft by departing employees, intellectual property leaks, security policy violations) using signals from M365, Entra, Defender, Windows endpoints, HR systems, and Adaptive Protection. Covers policy templates (data theft by departing users, data leaks, security policy violations), HRIS connector setup, indicator selection, sequence detection, anomaly detection, alerts triage, case investigation with content explorer, integration with eDiscovery and Communication Compliance, Adaptive Protection's automatic DLP policy adjustment, privacy controls (pseudonymization), role separation, and tenant-allow-list. WHEN: insider risk management, IRM policy, data theft departing user, IP leak detection, HRIS connector Purview, Adaptive Protection, insider risk indicators, sequence detection Purview, IRM case investigation, IRM privacy controls, IRM pseudonymization, insider risk Sentinel. DO NOT USE for DLP policy authoring (use purview-dlp-policy / purview-advanced-dlp), eDiscovery only (use purview-ediscovery), or generic insider risk strategy (use insider-risk-baseline).
Performs comprehensive code reviews for Python files following PEP 8 and Google Python Style Guide standards. Checks code quality, best practices, security, performance, maintainability, and style compliance. Use when reviewing Python code or when asked to check, audit, or improve Python code quality.
Use Responder for LLMNR/NBT-NS/MDNS poisoning and NTLMv2 hash capture. Use in a Windows domain network when passive credential capture or a man-in-the-middle attack is needed. Responder listens for name-resolution broadcasts (LLMNR/NBT-NS/MDNS) on the network and forges responses to induce targets to send NTLMv2 authentication hashes. Captured hashes can be cracked offline with hashcat or relayed to other services with ntlmrelayx. Use this skill for LLMNR poisoning, NBT-NS poisoning, WPAD proxying, NTLMv2 capture, or man-in-the-middle attacks.
(forwward) Performs paranoid code review checking trust boundaries, data integrity, performance, race conditions, error handling, and OWASP security. Triggers on code review, pre-merge checks, security audit, bug hunting, or any request to review, audit, or check code quality.
Use when hunting for exposed secrets and auditing how an app you own manages them — API keys, tokens, passwords, and private keys in source, git history, binaries, logs, config, and CI; plus vaulting, rotation, and least-privilege of the secrets themselves.
Use spray for high-performance directory brute-forcing and fingerprinting. Use when you need directory/file enumeration, backup file discovery, or fingerprinting of web targets. spray is a next-generation directory brute-forcing tool by chainreactors with more than 50% better performance than ffuf, supporting smart filtering, mask dictionaries, resume from checkpoint, and bulk targets. Consider this skill for any scenario involving directory brute-forcing, path enumeration, backup file scanning, or web fingerprinting
Postgres best practices maintained by Supabase, for Postgres running anywhere. Load this skill BEFORE writing or changing anything that lives in a Postgres database: creating or altering tables and columns (including choosing column types), schema design, migrations and declarative schema files, RLS policies and the tests that verify them, indexes, triggers, database functions, queues and scheduled jobs (pg_cron, pgmq), vector/semantic search (pgvector), and restoring dumps (pg_restore) or importing data. Also load it when diagnosing slow queries, high CPU, timeouts, EXPLAIN plans, connection exhaustion, locking, bloat, or rows visible to the wrong user or tenant. This is not just a performance guide — schema, migration, security, and SQL authoring tasks need these rules too, even for a one-column change or a single query.
Configure and manage Supabase projects using MCP (Model Context Protocol). Use this skill when working with Supabase databases, setting up MCP servers, designing database schemas, implementing Row Level Security, managing migrations, or building modern data architectures with PostgreSQL. Essential for Supabase development, database design, and AI-powered database operations.
Use for Teleport access list work with tctl; listing available resources (servers, databases, apps, etc.) before choosing an access grant target, creating new access lists, updating existing lists, or deleting/retiring lists. Trigger for requests like “give alice access”, “create an access list”, “standing access”, “Access Request”, “custom access list”, “show/list apps”, “show AWS IC permission sets”, “add AWS IC to this list”, “change owners/members/access”, “remove access”, “delete/retire/tear down an access list”. Handles access-type lists where Teleport creates supporting roles, and custom lists that grant existing roles/traits.
Twingate GraphQL API, CLI tools, and automation. Load when the user wants to automate Twingate via the API, write scripts against the GraphQL endpoint, generate or manage API tokens, use the Twingate CLI tools, or build automation pipelines. Also trigger on 'GraphQL', 'X-API-KEY', 'Twingate API', 'api/graphql', 'service account key', 'connector token provisioning', 'rate limiting', or any Twingate admin API mention. Also activate for: AWS tag-driven resource automation (EC2/ECS/RDS tag sync, Lambda, EventBridge/CloudWatch Events), a web UI for the Twingate CLI (tgcli-web, FastAPI, Docker/Kubernetes), the Twingate GitHub Action for CI runners, Jupyter/Python API tutorials, or generating a network health/insights report from a Network Events export.
Twingate Internet Security — DNS filtering, exit networks, browser security, and DNS-over-HTTPS. Load when the user mentions DNS filtering, content filtering, internet security, exit networks, egress routing, browser security, NextDNS, DoH, DNS categories, or profile priority. Also trigger on 'Internet Security', 'DNS Security Profile', 'fixed egress IP', or 'SaaS allowlisting' in a Twingate context. Also activate for DNS conflicts and symptom-shaped DNS queries: a third-party DNS filter or AV product conflicting with Twingate (Cisco Umbrella, DNSFilter, AdGuard, Avast Real Site Protection), CGNAT range conflicts (`100.96/12`), multiple-network-interface DNS resolution problems, "all nameservers have failed", DNS request delays on Linux, a Docker container's second DNS query returning the wrong IP, or a personal/DIY exit-network VPN built on Twingate.
Use when the user asks about IdP integration, SCIM provisioning, security policies, device trust, groups, users, or access control in Twingate. Activate for: SAML, SCIM, Okta, Entra ID, Google Workspace, JumpCloud, OneLogin, Keycloak, device trust, device posture, MFA enforcement, groups, JIT access, ephemeral access, auto-lock, offboarding, deprovisioning, multi-IdP deployments, or security policy configuration. Also activate for identity automation tooling: automating device trust from an MDM or EDR inventory (Jamf, Kandji, Intune, CrowdStrike, SentinelOne, FleetDM, Automox, JumpCloud, Mosyle, Datto RMM), migrating group access between IdPs, self-service or Slack-based group access requests, and location-based group switching.
Guides authorized web application and API security testing—scoping and rules of engagement, OWASP-oriented testing (injection, auth/session, access control, SSRF, XSS, CSRF, business logic), REST and GraphQL API security, Burp/ZAP-style manual methodology without requiring commercial tools, evidence and remediation reporting, and retest validation. Emphasizes written authorization and safe boundaries. Use for web pentest, OWASP web assessment, web app security test, API pentest, Burp-style testing, XSS or SQL injection testing when authorized—not network/AD/infra pentest (network-pentester), general multi-domain pentest orchestration (penetration-tester), LLM or agent adversarial testing (ai-redteam), enterprise adversary simulation or purple-team campaigns (red-team-specialist), SOC alert triage (soc-analyst), incident command (incident-responder), or CI/CD security gates and SBOM programs (devsecops).
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.