Kittysploit-framework is a Python exploitation framework for authorized penetration testing and red-team work that keeps scanning, exploits, command-and-control, proxy interception, and post-exploitation inside one console, aimed at security engineers and offensive-security operators who run web and API engagements.

What it is

Kittysploit-framework is an offensive-security console written in Python 3.9+ and released under the MIT licence. It bundles a modular exploitation console, a built-in command-and-control server, a proxy interceptor, workspaces with engagement scope, post-exploitation tooling, and an autonomous agent driven by a local large language model, with 8,000+ scanners, exploits, and post modules available from the same prompt. The project describes its own scope as "From recon to shell — in one console," and lists CLI, RPC, REST API, and MCP as automation surfaces alongside its marketplace of extensions for proxy, OSINT, GUI, and protocol work.

The concrete problem it solves is tool fragmentation. Most stacks force an operator to jump between a separate scanner, an exploit framework, a C2, a proxy, and a notebook, carrying scope and findings between them by hand. Kittysploit-framework replaces that chain by holding modules, sessions, scope, and agent plans in a single console, so an engagement's boundaries, hosts, and findings stay organised per job rather than scattered across five tools. It lives in the Python offensive-security ecosystem and is distributed from source, with documentation in the repository README and USAGE.md and a project site at kittysploit.com.

Key capabilities

• Modular console with search, use, set, and run across scanner, exploit, auxiliary, and post modules, covering 8,000+ modules.
• Built-in C2 with listeners, payloads, live sessions, pivots, and post-exploitation in the same console as the modules.
• Autonomous agent run against a local Ollama model, configured with --llm-local and --llm-model llama3.1:8b, requiring no cloud API key.
• Mission profiles such as safe-web, owasp-web-parallel, internal-lab, and bug-bounty-safe.
• Risk gates that block intrusive action unless explicitly approved: --approve-risk intrusive is required to run exploits, while --plan-only and --dry-run preview without live actions, and --approve-post-exploit enables optional read-only collection after a session.
• Evidence-gated exploit handoff and parallel specialists split by OWASP class, including injection, XSS, SSRF, auth, and authz.
• Workspaces and scope per engagement, plus repeatable workflows and playbooks, and a marketplace of extensions delivered through proxy, OSINT, GUI, and protocol add-ons.

Who uses it and how

• Bug-bounty hunters running the bug-bounty-safe profile, where the safe and scope-aware defaults suit externally hosted targets.
• Internal red teams working authorised lab networks, for example agent http://192.168.56.10 --profile internal-lab --goal obtain-shell --shell-hunter.
• Web and API testers who want OWASP-class coverage in parallel rather than sequentially, using --profile owasp-web-parallel.
• Automation and CI-style operators who drive the framework from CLI, RPC, REST API, or MCP instead of an interactive console.
• Post-exploitation operators who escalate from an interactive session into pivoting and collection using --approve-post-exploit.

Getting started

Clone the repository and run ./install/install.sh on Linux or macOS, then python3 kittyconsole.py; Windows users run install\install.bat followed by python kittyconsole.py. A one-line installer is also published: curl -fsSL https://raw.githubusercontent.com/SIA-IOTechnology/kittysploit-framework/main/install/install-standalone.sh | bash.

How it compares

No paid product and no comparable tool is named in the facts for this entry, so the framework stands alone in this registry. The README positions it against a general stack of separate scanner, exploit framework, C2, proxy, and notebook rather than against any named alternative.

When to use it — and when not to

A self-hoster must operate the Python 3.9+ environment, install the framework from source or the one-line installer, and run a local Ollama model to use the autonomous agent, since no hosted or managed option is described. Teams that want a fully managed service, or that cannot run a local LLM, should look elsewhere; intrusive and post-exploitation actions are deliberately gated behind explicit approval flags, which makes the tool a poor fit for anyone expecting unattended exploitation. Documentation is limited to the README and USAGE.md, and the repository shows 9 open issues against 624 stars and 98 forks, so operators should expect to read source and community channels such as the project Discord.

# Web/API mission (lab)
kittysploit> agent https://lab.local --profile owasp-web-parallel

# Go for a shell (authorized lab only)
kittysploit> agent http://192.168.56.10 \
  --profile internal-lab \
  --goal obtain-shell \
  --approve-risk intrusive \
  --shell-hunter

git clone https://github.com/SIA-IOTechnology/Kittysploit-framework.git
cd Kittysploit-framework
./install/install.sh
python3 kittyconsole.py

curl -fsSL https://raw.githubusercontent.com/SIA-IOTechnology/kittysploit-framework/main/install/install-standalone.sh | bash

git clone https://github.com/SIA-IOTechnology/Kittysploit-framework.git
cd Kittysploit-framework
install\install.bat
python kittyconsole.py

kittysploit> doctor
kittysploit> scanner -u http://192.168.56.10
kittysploit> search wordpress
kittysploit> agent https://lab.local --profile owasp-web-parallel --plan-only
kittysploit> agent http://192.168.56.10 --profile internal-lab --goal obtain-shell --approve-risk intrusive --shell-hunter

# Plan only (safe preview)
kittysploit> agent lab.local \
  --llm-local \
  --llm-model llama3.1:8b \
  --profile owasp-web-parallel \
  --plan-only

# Obtain a shell (lab / authorized targets only)
kittysploit> agent http://192.168.56.10 \
  --profile internal-lab \
  --goal obtain-shell \
  --approve-risk intrusive \
  --shell-hunter \
  --llm-local \
  --llm-model llama3.1:8b