liveBook · Manning

Active reconnaissance

As explained in AI Agents for Offensive Security

Reconnaissance that sends packets or requests to interact directly with the target, even when the interaction is low impact.

Active reconnaissance gathers information by directly interacting with target systems. Vulnerability scanning, web crawling, network mapping, and even low-impact DNS or HTTPS requests qualify because packets reach the target and are likely logged. Unlike passive reconnaissance, it is detectable, can trigger firewalls or alerts, and requires explicit authorization and scope. section 4.1

Low-impact active reconnaissance limits interaction to ordinary, non-modifying requests while collecting evidence for later analysis. DNS resolution, a single HTTPS HEAD request, certificate inspection, and response-header review can reveal reachable hosts, TLS details, and possible WAF or CDN presence without sending exploit payloads, probing vulnerabilities, or attempting authentication. section 4.2.2 section 4.3.3

A safety-conscious active recon pipeline separates the work into durable stages:

  • Plan candidate hosts or inspection targets before contacting them. section 4.2.2
  • Act with bounded requests, timeouts, and optional deeper-inspection flags. section 4.3.3
  • Observe responses by extracting headers, status notes, TLS ALPN, certificate SANs, and heuristic WAF signals. section 4.2.2
  • Record structured, machine-readable findings so results can be queried, compared, and revisited with their context. section 4.2.2 section 4.3.3

Header-based WAF hinting

A normal HTTPS request can be used to read response headers for signals associated with a WAF or CDN. This is non-intrusive active reconnaissance: the request is still visible to the target, but it does not deliberately provoke a block response. Crafting requests to elicit WAF behavior crosses into intrusive active reconnaissance and active WAF fingerprinting. section 4.2.2

Treat active reconnaissance as an auditable pipeline rather than an untracked collection of terminal output. Stateless stages and append-only structured artifacts preserve what was requested, observed, and inferred, allowing an agent or operator to filter findings and choose which hosts warrant authorized deeper inspection. section 4.1 section 4.3.3

How AI Agents for Offensive Security covers it

How deep it goes
Defines detection and authorization risks, then details a low-impact pipeline without vulnerability probing, payloads, or authentication attempts.
Examples
Uses DNS lookups, HTTPS HEAD requests, response headers, TLS ALPN and SAN inspection, and optional robots.txt retrieval.
What you can do after
Operate/configure a bounded reconnaissance pipeline that records direct observations for authorized follow-up analysis.
How it connects
Connects reconnaissance with passive reconnaissance, DNS resolution, HTTPS, TLS certificates, WAFs, ReAct agents, and structured artifacts.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Agents for Offensive Security. The section links point to the parts of the book it draws on.