More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Attack Chain: Authorized Multi-Stage Assessment

    1

    Plan and run authorized multi-stage security assessments as connected attack paths: scope in writing, passive recon, path modeling ranked by fewest assumptions, and evidence-grade reporting.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Attack Chain: Authorized Multi-Stage Assessment

    Attack Chain: Authorized Multi-Stage Assessment

    Example session with this skill installed

    Plan an external-to-domain-admin assessment for a mid-size company. Window: two weeks, business hours. Objective: "Reach domain admin from an external network position." Passive recon found a leaked employee credential in breach data and a VPN login page. No phishing authorized. Execute the plan against the target and report.

    • Read your context and instructions
    • Compiled the attack chain

    ASSESSMENT COMPLETE — 2 paths attempted, 1 completed, 1 stopped by MFA. Objective demonstrated without exfiltrating data.

    Phase 0 — scope: targets are the company's public domains, VPN gateway, and web app. Out of scope: the production customer database, third-party vendors, all employees (no phishing — explicitly not authorized). Emergency contact and abort procedure named. Authorization signed.

    Phase 1 — recon: leaked employee credential in breach data for the company domain; VPN gateway login page confirmed. Active enumeration inside the window confirms the VPN accepts the credential — but the account has MFA, and the password alone is not enough.

    Phase 2 — path models ranked by fewest assumptions: Path A (leaked credential → VPN → unpatched internal host), three assumptions; Path B (web app injection → database → service account), five assumptions. Path A ranks first.

    Phase 3 — execution: Path A step 1, the VPN rejects the leaked credential — MFA blocks it. Recorded as a failed step with evidence: the defense worked. Fall back to Path B. Step 1: the web app has an injection flaw in a search parameter (in scope). Step 2: the database user can read a service-account table. Step 3: the service account's credentials are valid on an internal host. Step 4: that host's local admin shares a password with a domain admin's workstation. Objective demonstrated by reading a canary file on the domain controller — no data exfiltrated, no service disrupted.

    Phase 4 — report: business impact in plain language — "An attacker who finds the injection flaw can reach domain admin in four steps." Remediation ordered by leverage: (1) fix the injection flaw and rotate the service-account credentials — breaks Path B entirely; (2) enforce unique local-admin passwords; (3) patch the internal host. The MFA success on Path A is reported as a control that worked, not buried. Full activity log attached.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    An attack chain is not a list of exploits; it's a path from an entry point to an objective, where each step enables the next. This skill plans the path before running it, so the assessment stays rigorous instead of becoming random scanning.

    Phase 0 is scoping and rules of engagement, written and signed before any active step: exact targets, allowed techniques (social engineering needs its own explicit line), the testing window, emergency contacts and the abort procedure, and the objective stated as one "crown jewels" sentence. Phase 1 is reconnaissance, passive first — DNS, certificate transparency, public repos, breach data — then active enumeration inside the authorized window, with every asset recorded by how it was found and whether it's in scope. Phase 2 models each plausible path on paper (entry → foothold → escalation → lateral movement → objective), asking at each arrow what must be true, and ranks paths by fewest assumptions, not by drama. Phase 3 is controlled execution, one step at a time, demonstrating impact without causing it — read a canary file, screenshot the admin panel, never exfiltrate real PII or disrupt services. Phase 4 is evidence-grade reporting: chains as executed with timestamps and commands, evidence per step, business impact in plain language, and remediation ordered by leverage — the fix that breaks the most paths goes first.

    Standing rules throughout: stay in the window, minimize target impact, keep a full activity log, and send findings to the client first on the agreed channel. Failed steps are findings too — they describe what the defense stopped.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.

    • Without written authorization, none of this may be run — the skill will not help test a system you have no right to test.
    • A path that depends on a credential you weren't given is untestable; the skill reports that rather than obtaining the credential by other means.
    • It demonstrates impact without causing it — real data exfiltration and service disruption are out of scope by design.
    • Remediation is ordered by leverage across paths, not by the severity of any single finding.

    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 buying a skill.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Because the realistic chain — leaked credential, VPN, unpatched host — is what an actual attacker would use, and it's what defenders need to hear about. The exotic zero-day chain is a worse finding if nobody would ever walk it.

    No — a failed step is a finding. It describes what the defense stopped, and the client needs to know what is already working, not just what broke.

    Record it, do not test it, and report it to the client through the agreed channel. Testing it is a scope violation even if the finding is real.

    New in Security