dreadnode
Dreadnode SDK
Decision gist · record as of 2026-08-14
Yes, with conditions. Install if you are building or testing AI security agents and need a unified framework for agent definition, evaluation, and red teaming. The active maintenance, low install friction, and comprehensive feature set (agents, evaluations, red teaming, tracing) make it a solid choice for security-focused AI workflows. However, verify the license before use in proprietary projects, budget for LLM API costs, and confirm that the red teaming attacks meet your research or production requirements. The large dependency footprint may be a concern in resource-constrained deployments.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.11 or later.
- Most use cases require API keys for LLM providers (OpenAI, Anthropic, AWS Bedrock) configured separately.
- Low install friction with a pure-Python wheel.
License · maintenance · safety
(unclear) — License treatment is unclear—no SPDX identifier or raw license text is available in the metadata. Verify the actual license before use in proprietary or copyleft-sensitive projects.
last release 2026-07-23 (22 days)
0 known vulnerabilities (OSV.dev, 2026-08-14) · 142,431 downloads/mo, #11,209 on PyPI
Alternatives
Verify before relying
pip install dreadnode
import dreadnode as dn
@dn.tool
def search_database(query: str) -> list[str]:
return ["CVE-2024-1234"]
@dn.agent(model="openai/gpt-4o", tools=[search_database])
def security_analyst():
"""You are a security analyst."""
trajectory = await security_analyst.run("Analyze vulnerabilities")
print(f"Steps: {len(trajectory.steps)}")- Whether the package's red teaming attacks (TAP, GOAT, Crescendo, AutoDAN-Turbo, ReNeLLM) are production-ready or research prototypes.
- Cost and rate-limit implications of running evaluations against large datasets with multiple LLM calls per example.
- Whether the local TUI runtime and platform sync work reliably without the Dreadnode platform backend.
- Performance characteristics when running agents with the full 44-dependency stack on resource-constrained environments.
What it is and what it does
Dreadnode is a Python SDK for building and testing AI security agents—systems that reason over multiple steps, call tools, and produce observable traces of their execution. It provides decorators to define agents with tools and hooks, run them against datasets with composable scorers, and trace all steps via OpenTelemetry. The package also includes a suite of AI red teaming attacks (TAP, GOAT, Crescendo, AutoDAN-Turbo, ReNeLLM) designed to probe LLMs for safety and security failure modes by systematically generating adversarial prompts.
The SDK integrates with HuggingFace for dataset loading, supports deployment via FastAPI or Ray, and includes a TUI for interactive development and platform sync. It depends on a large ecosystem: boto3 and AWS SDKs for cloud services, litellm for LLM routing, fastapi for serving, optuna for hyperparameter tuning, and data libraries like pandas and numpy. Most workflows require external LLM API keys (OpenAI, Anthropic, AWS Bedrock) and will incur costs proportional to token usage.
Use it for
- Build a multi-step security analysis agent that searches vulnerability databases and reports findings with automatic tracing.
- Run systematic red teaming attacks against an LLM to find jailbreaks or unsafe outputs using TAP or Crescendo.
- Evaluate a penetration-testing agent against a dataset of web applications with custom scorers measuring thoroughness and safety.
- Deploy an AI agent as a FastAPI service with built-in observability and real-time scoring hooks.
- Experiment with language-adapted red teaming by translating attack prompts or framing them in benign contexts.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, with conditions.
Install if you are building or testing AI security agents and need a unified framework for agent definition, evaluation, and red teaming. The active maintenance, low install friction, and comprehensive feature set (agents, evaluations, red teaming, tracing) make it a solid choice for security-focused AI workflows. However, verify the license before use in proprietary projects, budget for LLM API costs, and confirm that the red teaming attacks meet your research or production requirements. The large dependency footprint may be a concern in resource-constrained deployments.
Install
dreadnode on PyPI
Before you install
Low install friction with a pure-Python wheel. Active maintenance (latest release 22 days old). Requires Python 3.11 or later. The 44 runtime dependencies include heavy stacks (boto3, fastapi, datasets, litellm, optuna) which will pull in substantial transitive dependencies.
Requires Python 3.11 or later. Most use cases require API keys for LLM providers (OpenAI, Anthropic, AWS Bedrock) configured separately.
License in practice
License treatment is unclear—no SPDX identifier or raw license text is available in the metadata. Verify the actual license before use in proprietary or copyleft-sensitive projects.
Quickstart
pip install dreadnode
import dreadnode as dn
@dn.tool
def search_database(query: str) -> list[str]:
return ["CVE-2024-1234"]
@dn.agent(model="openai/gpt-4o", tools=[search_database])
def security_analyst():
"""You are a security analyst."""
trajectory = await security_analyst.run("Analyze vulnerabilities")
print(f"Steps: {len(trajectory.steps)}")Verify before relying
- Whether the package's red teaming attacks (TAP, GOAT, Crescendo, AutoDAN-Turbo, ReNeLLM) are production-ready or research prototypes.
- Cost and rate-limit implications of running evaluations against large datasets with multiple LLM calls per example.
- Whether the local TUI runtime and platform sync work reliably without the Dreadnode platform backend.
- Performance characteristics when running agents with the full 44-dependency stack on resource-constrained environments.
Package facts
| License | Not declared unclear |
| Python support | Supports the current Python release <3.14,>=3.11 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 44 packagesaiofilesartaws-sdk-bedrock-runtimeawscrtboto3coolnamecronitercycloptsdatasetsddgsfastapifastmcpfsspecgepahttpxjsonpath-ngjsonreflitellmlogfirelogurumarkdown-it-pymarkdownifymcpnetworkxnumpyoptunaorjsonpackagingpandaspillow |
| Maintenance | Actively maintained 22 days since the last release |
| First released | |
| Downloads | 142,431 / month, #11,209 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
Evidence: dreadnode-2.0.38-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “red teaming tools for llms”
- dreadnodeDreadnode is an SDK for building, testing, and evaluating AI security…
- deepteamDeepTeam is an open-source red teaming framework that simulates…
- pyritPyRIT is a framework for security professionals to identify and…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also deepteam · strands-agents-evals · bingo-ai · judgeval · langwatch-scenario · pyrit · nemo-gym · trinity-agent · agentops · opentelemetry-instrumentation-openai-agents