T3MP3ST

T3MP3ST is an offensive security framework that converts AI coding agents into autonomous tools for vulnerability research and exploitation.

T3MP3ST acts as an offensive-security war machine that coordinates with AI coding agents to perform reconnaissance, exploitation, and reporting tasks. It functions by bolting an arsenal onto a user's existing AI coding agent, such as Claude Code, Codex, or local models via Ollama and vLLM. Users define a target, and the framework executes the kill chain autonomously from a browser-based War Room or command-line interface. It emphasizes reproducibility by allowing users to recompute benchmark claims with a single command. The framework targets web applications, Capture The Flag challenges, and open-source software vulnerability hunting.

T3MP3ST at a glance

Best for: Software engineer
Runs on: Web, macOS, Windows, Linux, CLI, self-hosted
Pricing: Free
Released: July 2026
Category: AI coding tool
Website: github.com/elder-plinius/T3MP3ST

Key features

Autonomous kill chain execution for reconnaissance and reporting
Keyless operation using existing local AI coding agents
Support for local, offline models including Ollama and vLLM
Reproducible benchmarks with verification commands
Browser-based War Room interface
White-box repository analysis with tree-sitter ingestion
Integration with command-line tools for offensive operations

Use it for

Conducting automated web application reconnaissance and exploitation
Solving hint-free Capture The Flag challenges
Performing coordinated vulnerability disclosure for open-source projects
Analyzing source code for security vulnerabilities
Testing cloud infrastructure for misconfigurations

Alternatives to T3MP3ST

P3 Stack — Native T3 orchestration with worktree threads and delegated subagents.
Mantis — Automates vulnerability discovery and patching using AI agents.
ARTEX — Runs autonomous multi-agent penetration testing with a dual-graph architecture.
PentAGI — It autonomously manages penetration testing workflows within a secure Docker sandbox.
AiSOC — Runs locally without API keys with bundled models and auditable reasoning ledgers.
Claude Red — It primes Claude with expert-level methodology for specific attack surfaces.
CyberStrike — Autonomous AI agents execute recon, web, network, and cloud penetration testing.
Ctxfw — Reduces AI context token mass while preserving code structure and dependencies.
Offense — It gives coding agents structured security evidence through CLI, MCP, and HTTP.
Raptor — It automates complex security research workflows through an integrated agent.
HexStrike AI — Runs 150+ professional security tools via AI agents.
DeepTeam — It automates penetration testing using adversarial LLM simulations.

All T3MP3ST alternatives