LLMMap

What LLMMap is

LLMMap is a Linux security testing tool for examining LLM-integrated applications for prompt injection. It discovers HTTP injection points, creates targeted prompts, and checks candidate findings for reliability. Its four prompt packs cover 227 prompt injection techniques across 18 attack families. A Generator LLM creates goal-aware prompts, while a Judge evaluates target responses; detection can also use heuristics and optional semantic similarity. Provide a target URL or a Burp Suite request export, marking an injection location with an asterisk. LLMMap supports injection through query parameters, request bodies, headers, cookies, and paths, and lists obfuscation methods such as base64, homoglyphs, leet speak, and language switching. Supported backends are Ollama, OpenAI, Anthropic, and Google; Ollama is the local default and needs no API key. Safe mode is enabled by default and restricts scans to low-risk prompt families. The project lists proxy support for traffic inspection and a dry-run mode that plans a scan without connecting to the target. LLMMap is self-hosted, free, and available under the MIT license. It requires Python 3.11 or later and an LLM backend. The project describes it for authorized security testing of systems you own or are explicitly authorized to test.

Who it suits

LLMMap suits security practitioners who need to assess prompt injection in LLM-integrated applications they own or have written authorization to test. It is intended for Linux and requires Python 3.11 or later plus an LLM backend.

Limits

LLMMap requires Python 3.11 or later and an LLM backend; cloud backends may require API keys. Adaptive TAP and out-of-band detection are reserved for future versions and inactive in the v1.0.0 scan pipeline.

Getting started

1. Open the LLMMap project on GitHub.
2. Prepare a Linux system with Python 3.11 or later.
3. Provide a URL or Burp Suite request export and mark the injection location with an asterisk.
4. Choose an LLM backend; Ollama is the local default, while cloud backends may require API keys.
5. Use dry-run mode to plan a scan without making a network connection to the target.

Verdict

Choose LLMMap if you need a free, self-hosted Linux tool for authorized prompt injection testing with multiple prompt families and backends. Look elsewhere if you need adaptive TAP or out-of-band detection in the v1.0.0 scan pipeline.