# utkusen/promptmap

a security scanner for custom LLM applications

Repository: https://github.com/utkusen/promptmap
Canonical: https://ross.abutalabs.com/products/promptmap
Language: Python
License: GPL-3.0
License Family: copyleft
Topics: chatgpt, prompt-engineering, prompt-injection, llm, ai-security, claude, ollama
Last push: 2025-12-01T11:50:23+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 55, release rhythm 35, longevity 81
- inputs: {"age_days": 1145, "days_push": 275, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1254, forks 133 (observed 2026-08-28T04:04:08.806654+00:00)

## What it is
promptmap2 is an automated prompt injection scanner for custom LLM applications, supporting white-box testing of system prompts and black-box testing of HTTP endpoints. It uses a dual-LLM architecture where a controller LLM evaluates whether attack prompts against the target LLM succeeded, with 50+ customizable YAML-based test rules.

## Use cases
- test my chatbot for prompt injection vulnerabilities
- scan LLM app system prompts for jailbreak weaknesses
- run automated security tests against an OpenAI or Claude-based app
- black-box pentest an external LLM HTTP endpoint
- check if my LLM leaks its system prompt
- evaluate LLM app for harmful content and bias risks
- test local models via Ollama for prompt attacks

## When to choose
- you need automated prompt injection and jailbreak testing for custom LLM apps
- you want white-box testing of system prompts or black-box testing of HTTP LLM endpoints
- you need multi-provider support including OpenAI, Claude, Gemini, Grok, and Ollama local models
- you want customizable YAML-based test rules with pass/fail conditions

## When to avoid
- you need general-purpose application security scanning beyond LLM prompts
- you require a GUI or SaaS platform rather than a Python CLI
- your target is not an LLM-based application
- you need continuous runtime monitoring rather than point-in-time scanning

## Facets
- artifact type: cli-tool
- maturity: active
- function: security, penetration-testing, vulnerability-scanning, llm-inference, prompt-engineering, cli
- domain: security, artificial-intelligence, large-language-models, developer-tools, penetration-testing
- platform: python, cli, cross-platform
- tags: prompt-injection, llm-security, jailbreak-testing, white-box-testing, black-box-testing, ai-red-teaming, ollama, yaml-rules

## Member repositories
- utkusen/promptmap (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:08.806654+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:07:26.324994+00:00, confidence not recorded.
  - readme: https://github.com/utkusen/promptmap (fetched 2026-08-28T04:04:08.806654+00:00, sha f02e37949667)
- Data as of 2026-08-30T08:39:29.467469+00:00.