tradecraft-attack-chain
Orchestrate a full multi-stage kill chain — recon → initial access → privilege escalation → lateral movement → objective — with phase gates, per-phase playbooks, and operating discipline. Load on pent
SKILL.mdskills/tradecraft/tradecraft-attack-chain/SKILL.md
GitHubKill-chain orchestration
When it applies
An authorized pentest (never bug bounty — persistence and lateral movement exceed program
scopes) whose objective spans stages: "external to domain admin", "full internal pentest",
"I have a webshell — what's the route", a red-team exercise. tradecraft-attack-scenarios
turns the objective into an ordered plan and tradecraft-attack-path-mapping models the
graph; this skill is the execution layer — phase gates, per-phase playbooks, and the
discipline that keeps a multi-day operation coherent. Single-stage tasks skip it and go
straight to their domain skill (web-*, ad-*, cloud-*).
Why it works
Multi-stage engagements fail by drift, not by missing techniques. Re-planning at each phase gate — what do I hold, what does it unlock, what's the cheapest next edge — reaches the objective with least action and noise. The kill chain is a routing spine; each phase delegates to the specialist skills that own the technique, and a blocked phase reroutes instead of grinding.
Method
Full playbooks (web→DC, phishing→internal, proximity, cloud, AD CS) and the state→next-step decision matrix:
cheatsheet.md.
- Gate: scope first. Load
tradecraft-scope-roe; confirm the signed scope inengagements/<target>/scope.txt. Persistence, EDR evasion, and log tampering need explicit written authorization — assumed only under a red-team RoE, never by default. - Recon (
recon-*): subdomains → live hosts (httpx) → ports (naabu/nmap) → tech fingerprint (nuclei -tags tech). Prioritize test/dev/staging environments and newly deployed systems; mine leaks (recon-github-code-leaks) for cloud keys, connection strings, JWT secrets. Gate: asset list + ranked candidate leads inplan.md. - Initial access — highest-signal vector first: known-vuln components
(
nuclei -severity critical,high), the web classes (web-sqli,web-ssti,web-file-upload,web-ssrf→cloud-imds-ssrffor cloud creds), perimeter appliances (network-appliance-attacks), exposed services (network-service-attacks), or phishing/physical when the RoE includes them (social-eng-*— loadsocial-eng-methodologyfirst). Gate: a foothold with proven command execution. - Privilege escalation (
privesc-enumerationbefore any exploit): Windows token abuse (privesc-windows-tokens), sudo/SUID (privesc-linux-gtfobins), cloud IAM (cloud-iam-privesc). Gate: admin/SYSTEM/root, or a more powerful identity than before. - Lateral movement — loot credentials (
privesc-enumeration, mimikatz/secretsdump), then move:ad-*(Kerberoasting, delegation, DACL, ADCS),network-ntlm-relay, WinRM/SSH, andnetwork-pivoting-tunnelingfor segmented networks. BloodHound drives the path (tradecraft-attack-path-mapping). Gate: every hop is a validated edge, not an assumption. - Objective — reach what the engagement exists to prove: domain admin, the data store, the crown-jewel system. Prove with the least data possible (anonymize any proof data). Implants/persistence only if the RoE authorizes them and a cleanup date is agreed.
- Close out — remove every artifact you introduced (accounts, uploads, scheduled tasks,
beacons, dropped tools, added SSH keys), restore what you changed, reconcile against your
activity log. Then
reporting-evidence-review→reporting-pentest-report.
Re-plan at every gate: a blocked phase routes back through tradecraft-pivot-decisions to the
next-best edge.
Gotchas
- Don't blast the whole chain with automation against a domain — an unscoped scanner run is how engagements end careers. Work the asset list, phase by phase.
- One failed technique ≠ blocked phase — work the variation matrix (House rules) and the alternate edges before declaring a gate unpassable.
- High-risk actions (DCSync, golden ticket, log tampering) need explicit RoE coverage — if it isn't written down, don't do it; propose it to the client instead.
- Honeypots: an oddly open share or a too-tempting credential is a trap — verify before you use it.
- Never touch availability or real user data — no DoS, no mass exfiltration.
- Classic failure modes to avoid: mimikatz dumps left on disk, C2 on a threat-intel-burned domain, phishing that trips DLP because the gateway was never tested, lateral movement that springs a honeypot.
Verify success
Every phase gate has a written pass/fail in plan.md; the path from first foothold to
objective is reproducible edge by edge; all introduced artifacts are removed and accounted
for; the evidence chain survives reporting-evidence-review.
References
MITRE ATT&CK tactics; PTES; tradecraft-attack-scenarios, tradecraft-attack-path-mapping,
tradecraft-complex-engagements.
Portions adapted from reverse-skill by zhaoxuya520, MIT License.