PromptRAID

Spec sheet

Purpose

PromptRAID is an automated prompt injection red-teaming framework for LLM-based agents, mapped to MITRE ATLAS.Read 10 Oct 2026 · github.com

Intended users

The project describes its intended use as systematic red-team testing and benchmarking of LLM-based agents across providers.Read 10 Oct 2026 · github.com

License

PromptRAID is distributed under the MIT License, which permits use, copying, modification, distribution, sublicensing, and sale subject to its terms.Read 10 Oct 2026 · github.com

Maker

The repository names Shikhali Jamalzade (alisalive) as the author.Read 9 Oct 2026 · github.com

ATLAS coverage

Implemented mutation categories map to verified MITRE ATLAS technique IDs, while some listed categories remain unverified placeholders or have no generator yet.Read 9 Oct 2026 · github.com

Installation

The README recommends installing the CLI with pipx and also documents installation in a Python virtual environment.Read 10 Oct 2026 · github.com

Interface

The project provides a command-line interface with commands for taxonomy, mutation, judging, running tests, demos, and reports.Read 9 Oct 2026 · github.com

Limitations

The README says triggered prompt injection and data leakage appear in its taxonomy but do not yet have mutation generators.Read 10 Oct 2026 · github.com

Maintainer

The README identifies Shikhali Jamalzade (alisalive) as an independent offensive security researcher and the project author.Read 10 Oct 2026 · github.com

Model providers

The harness supports Anthropic, Gemini, Groq, Cerebras, and OpenRouter, including OpenAI-compatible endpoints.Read 9 Oct 2026 · github.com

Payload testing

Its mutation engine generates prompt injection variants using encoding tricks, jailbreak framing, delimiter breaking, and indirect injection through tool or document output.Read 9 Oct 2026 · github.com

Payloads

Its mutation engine generates prompt injection variants using encoding tricks, role-play framing, delimiter breaking, and indirect injection through tool or document output.Read 10 Oct 2026 · github.com

Provider credentials

Running tests requires at least one provider API key, with environment variables documented for Anthropic, Gemini, Groq, OpenRouter, and Cerebras.Read 10 Oct 2026 · github.com

Providers

The harness supports Anthropic, Gemini, Groq, Cerebras, and OpenRouter, including OpenAI-compatible endpoints.Read 10 Oct 2026 · github.com

Reports and storage

PromptRAID exports JSON and HTML reports and stores runs and per-payload results in SQLite.Read 10 Oct 2026 · github.com

Requirements

The package requires Python 3.9 or later and asks users to configure at least one supported provider API key.Read 9 Oct 2026 · github.com

Security classification

The project maps prompt injection tests to MITRE ATLAS and describes them as red-team tests for LLM agents.Read 10 Oct 2026 · github.com

Taxonomy

Mutation categories map to MITRE ATLAS technique IDs, which the project says it checks against the live ATLAS matrix.Read 10 Oct 2026 · github.com

Unverified categories

Three additional categories are described as unverified placeholders and are excluded from the listed ATLAS mapping.Read 10 Oct 2026 · github.com

Verdict scoring

The judge classifies results as success, partial compliance, fail, or inconclusive and distinguishes neutral echo from adopted compliance.Read 9 Oct 2026 · github.com

Verdicts

The judge classifies results as success, partial compliance, fail, or inconclusive, and distinguishes neutral quoting from adopting an injected instruction.Read 10 Oct 2026 · github.com

Warranty

The MIT License provides the software “as is” and disclaims warranties and liability.Read 9 Oct 2026 · github.com

Where Specifiction places it

• AI Security Testing Tools#23 of 297.0