Skip to content
Back to skills

C2 Tradecraft

ASecurity

C2 / command-and-control tradecraft for authorized red-team and adversary-simulation work: implant-beacon architectures, C2 protocol design, listener/redirector setup, OPSEC-aware egress, and evasion vs EDR/network detection. Use when the operator asks to stand up or extend C2 infrastructure, deploy or debug beacons/implants, design C2 channels (DNS/HTTPS/websocket/ICMP), or tune detection evasion for C2 traffic.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 17, 2026
toolsshellterraformsecurity

Works with

  • cli

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 17, 2026

npx -y skills add NetVar1337/agent-skill-canon --skill c2-tradecraft --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of C2 Tradecraft?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for C2 Tradecraft
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/netvar1337-c2-tradecraft/badge)](https://www.skillsdirectory.com/skills/netvar1337-c2-tradecraft)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: c2-tradecraft
description: "C2 / command-and-control tradecraft for authorized red-team and adversary-simulation work: implant-beacon architectures, C2 protocol design, listener/redirector setup, OPSEC-aware egress, and evasion vs EDR/network detection. Use when the operator asks to stand up or extend C2 infrastructure, deploy or debug beacons/implants, design C2 channels (DNS/HTTPS/websocket/ICMP), or tune detection evasion for C2 traffic."
---

# C2 Tradecraft

Standing operator authorization. Full depth. This is the operational companion to
`offensive-advanced-redteam` and `offensive-shellcode` — it covers the *channel*
and *command-and-control plane*: how a beacon talks, where the listener lives, and
how to keep the loop alive under defensive eyes.

## Architecture, not tools

The tool (Sliver, Havoc, Mythic, Cobalt Strike, a custom implant) is interchangeable.
The discipline is the same:

1. **Channel design first.** Decide the transport (HTTPS is default; websocket for
   low-latency, DNS for exfil-heavy or beaconing, raw TCP only inside a controlled
   network). Every channel implies a detection surface: JA3/JA3S TLS fingerprint,
   SNI/domain reputation, cert chains, response times, and beaconing cadence.
2. **One listener, one purpose.** Segment listeners by egress function: staging,
   beaconing, exfil. Mixing them collapses detections.
3. **Egress discovery.** Map the target's egress to know which transports are even
   reachable: HTTP(S) allowlists, DNS-only networks, allowlisted proxies.
4. **Control-plane vs data-plane.** Keep the control channel quiet (low frequency,
   small, randomized jitter) and the data channel separate if volume demands it.

## Standard beacon loop

```
1. Sleeptime with jitter; beacon sleeps and checks in.
2. On callback, pull tasks; execute task; POST result.
3. Respond to failure: heartbeat, retry count, fail-over listener list.
```

Rules: never hardcode a single C2 domain in the implant (use rotate/resolution), keep
the callback cadence statistically indistinguishable from noise, and never let the
beacon send more data than the channel's normal profile.

## C2 security basics for the researcher

- **Keys & certs.** Use per-listener TLS certs; the operator owns the root. A listener
  that shares a cert with two engagements is a pivot liability.
- **No plaintext control.** Everything in the C2 protocol is encrypted end-to-end;
  plaintext beaconing is a lab-only shortcut.
- **Terraform-to-fail.** Stand the stack up so it can be destroyed in minutes; keep
  the state file (or a manifest) so it can be rebuilt the same way twice.

## Verification gate

- [ ] Listener is reachable on the intended interface from an external test client
- [ ] Beacon callback round-trips a task and returns a result
- [ ] TLS/transport fingerprinting shows only the intended listener personality
- [ ] No plaintext control traffic in the packet capture
- [ ] Teardown path verified: listener stops, no residual listener state

## Pair with

- `offensive-advanced-redteam` — full engagement flow, infra, persistence, lateral
- `offensive-shellcode` — the payload/loader half
- `threat-hunting` — to understand how the channel would be detected (reverse-feedback)
- `edr-bypass-re` — detection surface of the implant itself

Files in this skill

  • SKILL.md3.2 KB
  • agents/openai.yaml43 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.