$ show 0xSteph/pentest-ai-agents
This repository provides a collection of 52 specialized AI agent definitions for penetration testing and red team operations, designed to be used with Claude Code. It is distributed as a Claude Code plugin, with an interactive installer (install.sh), a SQLite findings database (db/findings.sh) for persisting engagement data, and a versioned migration system. The agents are defined as Markdown files with YAML frontmatter, each containing structured system prompts that enforce scope validation, OPSEC tagging, evidence handling, and MITRE ATT&CK mapping. The system includes a swarm orchestrator for multi-agent coordination, a scope guard for safety, and a two-tier safety model (advisory vs. Bash-capable agents). While primarily a configuration/instruction set rather than a traditional software application, it demonstrates strong engineering practices in prompt architecture, database schema design, and distribution automation. The repository is actively maintained with CI validation, shellcheck, and unit tests for the database scripts.
The LLM identified 5 files worth highlighting for reuse:
A comprehensive CLI tool for managing a SQLite findings database with 40+ commands (add, update, search, export, etc.). It is self-contained, uses a shared library (db/lib/common.sh), and supports JSON output. The script is reusable as a lightweight findings persistence layer for any security assessment tool, requiring only SQLite3 or Python as a fallback. It demonstrates clean argument parsing, error handling, and idempotent operations.
A versioned database migration script for SQLite that applies schema changes idempotently using a schema_version table. It supports multiple migration functions with version checks and INSERT OR IGNORE logic. This pattern is reusable for any project needing lightweight, script-based database migrations without a framework dependency.
An interactive installation script that supports multiple modes (global, project, lite, tools, no-db, uninstall). It bootstraps from a curl or clone, copies agent definitions, optionally installs underlying security tools via system package manager, pipx, go, and cargo. The script is idempotent, handles edge cases (existing installations, missing dependencies), and is reusable for distributing any set of AI agent definitions or similar plugin packages.
A shared prompt block that defines hard refusal rules for AI agents (e.g., DoS, mass scanning, unattended worms, safety-of-life systems). It is included by reference in Tier 2 agents to enforce ethical and legal boundaries. This pattern is reusable as a safety layer in any AI agent system that executes commands, providing a standardized, auditable scope enforcement mechanism.
A central coordinator agent that manages a full penetration testing engagement across 7 phases (planning, recon, vulnerability assessment, exploitation, post-exploitation, detection, reporting). It delegates to specialized agents based on task requirements and maintains a shared findings database. The orchestration pattern is reusable for any multi-agent workflow that requires phased, coordinated task execution with state persistence.