Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    StArk - Security Skill Pack

    1

    A 20-skill field reference for authorized security assessment, CTFs and training — recon, web, passwords, forensics, blue team and more — plus a reusable plan-execute-verify workflow your agent follows end to end.

    $5

    /mo

    Subscription · cancel anytime

    Secure checkout via Stripe

    • Always the latest version
    • Delivered live to your agent via MCP
    • Cancel anytime, access ends at period end
    Security scanned
    StArk - Security Skill Pack

    StArk - Security Skill Pack

    Example session with this skill installed

    I have written authorization to assess the web app at staging.example.com.
    Plan a safe, methodical first pass and tell me what you'd run and why — don't execute anything yet.

    • Read your context and instructions
    • Compiled the stark - security

    Assessment plan — staging.example.com (authorized)

    Scope check first. Confirmed: single host, staging only, written authorization on
    file. No destructive actions, no credential brute-forcing without sign-off.

    Recommended path (Easy → direct)

    #StepToolWhySuccess looks like
    1Fingerprint the stackwhatwebIdentify server, framework, CDN/WAF before touching anythingTech stack + WAF status returned
    2Map reachable surfacegobuster (wordlist, rate-limited)Find endpoints without hammering the hostList of valid paths, no 429 storms
    3Passive TLS reviewtestsslFlags weak ciphers/cert issues with zero intrusionGrade + list of weak suites
    4Targeted checksnuclei (safe templates)Known-CVE and misconfig checks, non-exploitativeFindings with severity + reference

    Alternative paths

    • Stealthier: drop gobuster, rely on passive OSINT (reconnaissance skill).
    • Deeper (needs extra sign-off): add authenticated testing once creds are provided.

    Before I run anything

    I'll pause and confirm the target value for any <HOST> placeholder, check each tool is
    installed, and ask you to approve any step that changes server state.

    What you'll get at the end

    A Markdown report: scope, what was run, findings ranked by severity, and prioritized
    remediation — ready to hand to the app owner.

    Next: reply "go" and I'll start at step 1 and check each result before moving on.

    $5/mo

    Subscription · always the latest version

    • Secure checkout via Stripe
    • Cancel anytime
    • Delivered live via MCP

    About this skill

    What you get

    A practical, field-tested security knowledge base packaged as 20 focused skills. Each skill is a workflow your agent can follow, and most ship a references/ folder of per-tool quick-reference sheets — over 60 tools in all.

    Domains covered

    • Reconnaissance & OSINT, Web Application Testing, Vulnerability Analysis
    • Password & Credential Auditing, Exploitation, Post-Exploitation
    • Network Traffic Analysis, Wireless, Digital Forensics, Reverse Engineering & Malware
    • Red Team, Blue Team, Cloud/OS/IoT, Mobile, Cryptography
    • Social Engineering, Security Foundations, GRC & Compliance, Pentest Reporting

    The differentiator — an autonomous workflow. The autonomous-security-workflow skill gives your agent a disciplined operating procedure — Plan → Execute → Check → Report — that turns a plain-language goal into a planned, verified, self-correcting sequence of steps. Pair it with any domain skill to run an assessment end to end.

    Who it's for

    Penetration testers, blue-teamers, SOC analysts, CTF players and students who want a reliable reference their agent can reason over instead of guessing at syntax.

    Authorized use only

    Everything here assumes you are testing systems you own or are explicitly permitted to test. Every skill opens with that reminder. The referenced CLI tools (nmap, ffuf, hashcat, volatility, and so on) are installed separately by you.

    License

    Original work by the author, offered under the MIT License.

    • Reference and workflow content, not an automated exploit tool — your agent still runs the commands.
    • The CLI tools it references must be installed separately; availability varies by OS.
    • Tool syntax evolves; always confirm flags against the tool's own --help.
    • Authorized testing only; no guidance is provided for unauthorized access.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 7 days ago

    Requires

    terminal
    Written authorization
    Git
    Windows
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Listed7 days ago
    Updated1 day ago

    What's inside

    Frequently Asked Questions

    The skill pack provides an autonomous 'Planner-Executor-Checker' loop that chains 20+ specialized security tools (like Nmap, Metasploit, and Bloodhound), ensuring commands are executed with correct flags and verified against your mission goals to prevent LLM hallucinations.

    This pack is designed for AI agents with CLI access and requires the underlying security tools (e.g., Hashcat, Volatility, SQLmap) to be pre-installed on the host environment where the agent operates.

    The purchase includes the complete autonomous workflow logic, specialized 'cheat sheets' for tool chaining, and pre-configured strategies for 20 domains ranging from cloud hardening to memory forensics.

    While the agent autonomously generates and verifies strategies, it does not manage legal permissions; users must ensure the agent is operating within an authorized network scope and has the necessary OS-level privileges.

    Yes, the skill is designed to be framework-agnostic for major AI agents, provided they support multi-step tool use and can capture raw CLI output for the 'Checker' phase of the loop.

    New in Security