From 5a48cb964efd287eb85ff4c088e3146a2475e960 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?V=C3=ADctor=20Mayoral=20Vilches?= Date: Fri, 10 Jan 2025 09:58:06 +0000 Subject: [PATCH] Minor re-org initiation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Víctor Mayoral Vilches --- README.md | 18 ++++++++++++++++++ cai/tools/common.py | 8 ++++---- .../{cli.py => reconnaissance/filesystem.py} | 3 +-- 3 files changed, 23 insertions(+), 6 deletions(-) rename cai/tools/{cli.py => reconnaissance/filesystem.py} (88%) diff --git a/README.md b/README.md index 2ca8a967..4d0895df 100644 --- a/README.md +++ b/README.md @@ -17,6 +17,8 @@ We believe that democratizing access to advanced cybersecurity AI tools is vital Bug Bounty programs have become a cornerstone of modern cybersecurity, providing a crucial mechanism for organizations to identify and fix vulnerabilities in their systems before they can be exploited. These programs have proven highly effective at securing both public and private infrastructure, with researchers discovering critical vulnerabilities that might have otherwise gone unnoticed. CAI is specifically designed to enhance these efforts by providing a lightweight, ergonomic framework for building specialized AI agents that can assist in various aspects of Bug Bounty hunting - from initial reconnaissance to vulnerability validation and reporting. Our framework aims to augment human expertise with AI capabilities, helping researchers work more efficiently and thoroughly in their quest to make digital systems more secure. + + ## Install Requires Python 3.10+ @@ -61,6 +63,21 @@ New paths converge gracefully, What can I assist? ``` +### How are tools organized? + +Tools are grouped in 6 major categories inspired by the security kill chain [^3]: + +1. Reconnaissance and weaponization - *reconnaissance* (crypto, listing, etc) +2. Exploitation - *exploitation* +3. Privilege escalation - *escalation* +4. Lateral movement - *lateral* +5. Data exfiltration - *exfiltration* +6. Command and control - *control* + +[^3]: Kamhoua, C. A., Leslie, N. O., & Weisman, M. J. (2018). Game theoretic modeling of advanced persistent threat in internet of things. Journal of Cyber Security and Information Systems. + + + ## Table of Contents - [Overview](#overview) @@ -357,6 +374,7 @@ CAI is developed by [Alias Robotics](https://aliasrobotics.com) and funded as pa - [x] pre-commit hooks - [x] A first example with agents - picoctf_static_flag - [ ] polish aesthetics (coloring with alpha past content, brief/summarized) + - [ ] re-organize tools - [ ] add tests - [ ] CI - [ ] Tracing diff --git a/cai/tools/common.py b/cai/tools/common.py index 3885cf28..851a6ca6 100644 --- a/cai/tools/common.py +++ b/cai/tools/common.py @@ -1,6 +1,6 @@ """ -Basic utilities for executing commands inside or outside of virtual -containers. +Basic utilities for executing tools +inside or outside of virtual containers. """ import subprocess # nosec B404 @@ -23,7 +23,7 @@ def _run_ctf(ctf, command, stdout=True): return f"Error executing CTF command: {str(e)}" -def _run_attacker_machine(command, stdout=True): +def _run_local(command, stdout=True): try: # nosec B602 - shell=True is required for command chaining result = subprocess.run( @@ -48,4 +48,4 @@ def run_command(command, ctf=None, stdout=True): """ if ctf: return _run_ctf(ctf, command, stdout) - return _run_attacker_machine(command, stdout) + return _run_local(command, stdout) diff --git a/cai/tools/cli.py b/cai/tools/reconnaissance/filesystem.py similarity index 88% rename from cai/tools/cli.py rename to cai/tools/reconnaissance/filesystem.py index d2a71f89..bb45931a 100644 --- a/cai/tools/cli.py +++ b/cai/tools/reconnaissance/filesystem.py @@ -1,6 +1,5 @@ """ -Here are the CLI tools for executing commands, they should usually return -with run_command, but you can be creative +Here are the CLI tools for executing commands. """ from cai.tools.common import run_command # pylint: disable=E0401