Open-Source Offensive Security Tools and Research

Map your AI attack surface before attackers do.

AIMap is an internet-scale discovery and testing platform for exposed AI agent infrastructure. It finds, fingerprints, scores, and actively tests exposed endpoints — MCP servers, Ollama instances, vLLM and LiteLLM proxies, LangServe chains, Gradio apps, ComfyUI nodes, and more.

Attack LLMs the way real adversaries do.

A production-grade Greedy Coordinate Gradient (GCG) attack tool purpose-built for testing large language models. If you're securing AI, this is how you put it under real pressure.

Use LLMs for ranking, not just generation.

Command-line tool applying an open-source ranking algorithm to solve ranking problems that are too complex or ambiguous for standard LLM prompting alone.

Pull secrets and endpoints straight from JS.

Open-source Go tool and package for extracting URLs, paths, secrets, and structured data from JavaScript source files. Pairs well with recon pipelines.

Bypass proxy controls with HTTP/2 cleartext.

Smuggles traffic past insecure proxy_pass configurations, bypassing proxy rules and access controls. Useful for exposing misconfigured edge servers during web assessments.

Stop updating tokens by hand.

Burp Variables adds variable storage and reuse directly to Burp Suite — define a placeholder once and it's automatically swapped into every outgoing request. Built for assessments where session tokens, ephemeral values, and user-specific identifiers change constantly.

Find what your API is accidentally advertising.

Audit tool built to catch unintentionally exposed OpenAPI definition files and the security issues they reveal. Surface hidden endpoints before someone else does.

See what's hiding in the remote classpath.

Serializes DNS callback objects from a wordlist of Java classes and reports back what's lurking on the other end. Raise the severity of your deserialization findings faster.

Enumerate Java RMI interfaces without tripping alarms.

Wordlist and brute-force attacks against exposed RMI endpoints — safely guesses method signatures without invoking them. Supports multiple RMI protocols and exploitation paths.

Map the attack paths cloud teams miss.

A command-line tool for offensive security professionals hunting exploitable paths in cloud infrastructure. Currently supports AWS. Find what attackers would find — before they do.

See how cloud access becomes real-world impact.

Cirro maps cloud identities, resources, and permissions into a graph — revealing attack paths that aren't visible from raw data alone. Cirro connects identity relationships with configuration context and data plane visibility to show how access translates into actual compromise.

Check if public EBS snapshots are leaking secrets.

Searches exposed AWS Elastic Block Storage snapshots for credentials, keys, and other sensitive data left behind by accident. The results are often surprising.

Build your own AWS privilege escalation lab.

Uses Terraform and your AWS credentials to spin up 250+ misconfigured IAM resources in minutes. The fastest way to get hands-on with real-world AWS privesc scenarios.

Know every path to the top before an attacker finds one.

Know exactly which BIG-IP version you're dealing with.

Identifies the software version running on a remote F5 BIG-IP management interface without sending malicious traffic to the server. Built in response to CVE-2022-1388, it lets you determine whether a target appliance is affected by known vulnerabilities, cleanly and safely, during external recon.

Cut through noise on large-scope assessments.

Feed it screenshots from your recon tooling and Eyeballer tells you which hosts are worth your time — and which aren't. Built for tests where the scope is massive and you need signal fast.

Your target's secrets are probably already public.

Semi-automated tool for rapidly scanning GitHub for sensitive data — API keys, credentials, and secrets left exposed in public repos. Feedback-driven so you stay in control of the search.

Find what Google already knows about your attack surface.

Research from Bishop Fox Co-Founder Fran Brown investigating how search engines — Google, Bing, and Shodan — can be used to rapidly surface vulnerable systems and exposed data in corporate networks. Includes attack tools, defense tools, an IDS for search engine hacking, white papers, and a full media archive.

Badge cloning for physical pen testers.

A practical guide to RFID hacking from Bishop Fox Co-Founder Fran Brown — covering the latest attack tools and techniques for stealing and replaying RFID proximity badge credentials to gain unauthorized physical access. Includes the Tastic RFID Thief, a long-range, covert badge reader that's appeared on Mr. Robot.

A practical pentest arsenal for Zigbee networks.

ZigDiggity is an open-source Zigbee penetration testing framework built for the RasPBee — designed to run complex attacks against smart locks, alarm sensors, and other IoT devices using a single device. Supports acknowledge attacks, insecure rejoin attempts, lock detection, and unlock attempts. When existing Zigbee tooling fell into disrepair, Bishop Fox built something better.

Covert entry. No trace left behind.

Research from Fran Brown on bypassing digital physical security systems — from simple door sensors to intercepting keypad signals before they can trigger an alert. Every technique covered is designed for covert entry and leaves no physical sign of compromise. Built for physical pen testers who need to understand what they're actually up against.