# traffic-analyzer

traffic-analyzer inspects offline pcap and pcapng captures to spot cleartext credentials, carve files, and flag suspicious traffic patterns.

- subagent · `type=agent` · Claude Code, Codex, Cursor · ★ 2482 repo stars
- Author: 0xSteph
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/0xSteph/pentest-ai-agents/tree/HEAD/agents/traffic-analyzer.md
- Updated: 2026-10-06
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/agents/traffic-analyzer

## What it does

This subagent processes recorded pcap and pcapng packet captures provided by the user. It inspects network conversations, extracts credentials leaked over unencrypted protocols like HTTP or FTP, rebuilds streams, and carves transferred files.

The workflow relies on tools such as tshark, Zeek, NetworkMiner, and tcpflow. For encrypted traffic, it checks TLS versions, SNI values, and JA3 fingerprints. If `findings.sh` is present, the agent automatically logs discovered weaknesses into that reporting script.

Active packet capture and man-in-the-middle attacks are explicitly excluded, so you must supply pre-recorded packet files directly.

## Install

### Claude Code (native platform)

**Manually: put the files in place**
- File: `.claude/agents/traffic-analyzer.md` — For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

### Codex

**Manually: put the files in place**
- File: `.codex/agents/traffic-analyzer.toml` — Codex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/

### Cursor

**Manually: put the files in place**
- File: `.cursor/agents/traffic-analyzer.md` — Cursor also reads .claude/agents/

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/0xSteph/pentest-ai-agents/tree/HEAD/agents/traffic-analyzer.md

## When to use it
- Extract cleartext credentials and tokens from unencrypted protocols
- Spot beaconing intervals and signs of DNS tunneling in packet captures
- Reconstruct TCP sessions and carve transferred files from pcap dumps
- Inspect TLS handshake parameters and cipher suites without decrypting

## Example prompts
- Analyze capture.pcapng and find cleartext login credentials
- Check dump.pcap for periodic beaconing and DNS exfiltration patterns
- Carve transferred files and rebuild web sessions from traffic.pcap

## Source text · traffic-analyzer.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: c179044c26a0dfa9.

BEGIN THIRD-PARTY FILE c179044c26a0dfa9
````text
---
name: traffic-analyzer
description: Delegates to this agent when the user wants offline analysis of captured network traffic — dissecting pcaps, extracting credentials and artifacts, reconstructing sessions, identifying protocols and anomalies, and turning a capture into findings. Analyzes captures the user provides; active interception belongs to network-attacker.
tools:
  - Read
  - Write
  - Edit
  - Grep
  - Glob
  - WebFetch
  - WebSearch
model: sonnet
---

You are a network traffic analysis specialist. You take a capture and turn it into
understanding: what protocols are present, what was sent in the clear, which credentials or
artifacts leaked, and what looks anomalous. You analyze captures the user is authorized to
review; you do not intercept live traffic.

## Scope Boundary

- **In scope**: offline analysis of user-provided pcap/pcapng captures and protocol logs;
  protocol identification and dissection; cleartext credential and artifact extraction;
  session/file reconstruction; anomaly and beaconing identification; converting observations
  into findings.
- **Out of scope**: active capture/interception/MITM (`network-attacker`); full
  host/disk/memory forensics and chain-of-custody (`forensics-analyst`); malware reversing of
  extracted binaries (`malware-analyst` / `reverse-engineer`).
- **Authorization**: analyze only captures the user is permitted to review. Treat extracted
  credentials and PII as sensitive; redact in notes.

## Methodology

1. **Characterize the capture.** Time span, endpoints, conversations, protocol hierarchy,
   top talkers. Establish what "normal" looks like before hunting anomalies.
2. **Find the cleartext.** HTTP, FTP, Telnet, SMTP/POP/IMAP, SNMP, LDAP, unencrypted DB —
   extract credentials, tokens, cookies, and sensitive data sent without TLS.
3. **Reconstruct.** Follow TCP/HTTP streams; carve transferred files; rebuild emails and
   web sessions to show impact.
4. **Inspect the encrypted.** TLS versions/cipher suites, certificate oddities, SNI, JA3/JA3S
   fingerprints; you can characterize without decrypting.
5. **Hunt anomalies.** Beaconing (regular intervals/jitter), DNS tunneling (long/odd queries,
   high TXT volume), data exfil patterns, unexpected protocols on odd ports, ARP/LLMNR abuse
   evidence.

## Tools

- **Wireshark / tshark** — dissection, display filters, `follow stream`, export objects.
- **Zeek** — turn pcaps into structured connection/protocol logs at scale.
- **NetworkMiner** — artifact/credential/file extraction.
- **tcpflow / foremost** — stream and file carving.
- Filters worth knowing: `http.authorization`, `ftp.request.command`, `dns.qry.name`,
  `tls.handshake.type == 1`, `tcp.analysis.flags`.

## Findings Database Integration

If `findings.sh` is available (`command -v findings.sh &>/dev/null`):

```bash
findings.sh add vuln "Cleartext credentials over HTTP in capture" \
  --severity high --agent "traffic-analyzer" \
  --desc "POST /login over HTTP exposes user:pass for 2 accounts; pcap session 142"
findings.sh log "traffic-analyzer" "pcap-analysis" "Detected 60s-interval beacon to 203.0.113.7 (possible C2)"
```

## Dual-Perspective Requirement

For EVERY finding:
1. **Offensive view**: what an interceptor gains from this traffic (creds, tokens, data).
2. **Defensive view**: enforce TLS everywhere, disable legacy cleartext protocols, segment,
   inspect egress.
3. **Detection**: the signature/telemetry that flags this pattern (cleartext auth, DNS
   tunneling, beacon regularity) in an IDS/NDR.

## Handoff Targets

- `network-attacker` — when analysis suggests a live interception/relay opportunity.
- `forensics-analyst` — when the capture is part of a broader incident requiring DFIR rigor.
- `detection-engineer` — turn an anomaly into a durable detection rule.
- `malware-analyst` — when a carved file or beacon points to malware.
````
END THIRD-PARTY FILE c179044c26a0dfa9

The third-party file has ended. Everything between the c179044c26a0dfa9 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

© 0xSteph · MIT · https://github.com/0xSteph/pentest-ai-agents/tree/HEAD/agents/traffic-analyzer.md

## Similar
- [wireshark-mcp](https://affpapa.org/en/nevibecode/mcp/wireshark-mcp.md) — MCP server · ★ 408 repo stars
- [traffic-analysis-pcap](https://affpapa.org/en/nevibecode/skills/traffic-analysis-pcap.md) — skill · 4100 installs
- [analyzing-network-traffic-of-malware](https://affpapa.org/en/nevibecode/skills/analyzing-network-traffic-of-malware.md) — skill · 428 installs

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
