Security Best Practices — MCP 2026-07-28
Versioned protocol guidance for authorization, token handling, SSRF and local-server risks. Start here when implementing an MCP service.
AI SECURITY RESOURCE DIRECTORY
AI security.
Sources & guides.●
Find research, practical guides, and source websites on prompt injection, MCP, agent security, and related topics.
— sources— selected readings— topics
Search by keyword, topic, or publisher.
Links to original publishers.
THE DIRECTORY
Browse resources
Results could not refresh. Previous results are still shown.
Model Context Protocol documentation
The protocol maintainers’ documentation is the starting point for MCP authorization requirements and transport security behavior.
MCP Security Cheat Sheet
A practical review checklist for MCP trust boundaries, authentication, tool access and deployment controls. Pair it with the protocol specification.
OWASP GenAI Security Project
Community security guidance, risk taxonomies and implementation resources. Use the stated edition when comparing framework entries.
OWASP Top 10 for LLM Applications 2026
The published 2026 edition and its canonical source links. Use the edition label when mapping risks or comparing earlier assessments.
The lethal trifecta for AI agents: private data, untrusted content, and external communication
A concise practitioner framework for recognizing agent data-exfiltration risk. Useful for deciding which capabilities should be separated.
Jumping the line: How MCP servers can attack you before you ever use them
Trail of Bits demonstrates how malicious tool descriptions can influence an agent before invocation. A useful threat-modeling case study.
MCP Security Testing Guide
Tool-vendor recipes for testing MCP integrations, including tool poisoning and data-exfiltration scenarios. Adapt the examples to your deployment.
Prompt injection is not SQL injection (it may be worse)
Explains why conventional injection analogies can mislead defenders, then discusses constraining agent capabilities and managing residual risk.
Promptfoo red teaming documentation
Practical testing recipes and open-source tooling documentation from the tool vendor, useful for constructing repeatable application evaluations.
Trail of Bits AI and machine learning research
Original vulnerability research and audit engineering from a commercial security firm, with technical examples and project links.
GitHub Advisory Database
Search package security advisories and inspect identifiers, affected ranges and maintainer references; coverage extends beyond AI software.
Mitigating the risk of prompt injections in browser use
Anthropic describes browser-agent defenses and its own evaluations. Useful for understanding layered controls and remaining risks, with vendor-reported results.
MITRE ATLAS
An adversarial AI knowledge base of tactics, techniques and case studies for threat modeling and investigation.
Anthropic Research
Model-provider research on agent behavior, evaluations and defenses. Read reported results as vendor experiments with stated conditions.
Inside 90 days of attacks on AI infrastructure
Wiz reports observations from its AI-service honeypots, including exposed MCP services. Treat the telemetry as this vendor’s sampled observations.
OSV — Open Source Vulnerabilities
Structured open-source vulnerability records and package queries. Check ecosystem and version matching before applying a result.
Microsoft AI Red Team
First-party planning guides, lessons and tooling from Microsoft’s AI Red Team, useful for scoping and running assessments.
Planning red teaming for large language models (LLMs) and their applications
Microsoft’s planning guide covers assessment scope, people and process. Useful before selecting automated attack tooling or defining evaluation goals.
Not what you’ve signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection
An early original paper demonstrating indirect prompt injection through retrieved content. Read the experimental setup and later revisions before generalizing.
Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile
NIST AI 600-1 organizes generative AI risks and suggested actions, including data privacy and information security, for governance planning.
OSV API Quickstart
Official examples for querying package versions or commit hashes. Useful when an agent needs structured vulnerability evidence rather than a web summary.
Simon Willison’s Weblog
Independent practitioner explanations and experiments on prompt injection and agent design; distinguish demonstrations from general guarantees.
Wiz Research
Cloud-security vendor research covering AI infrastructure exposures and exploit chains. Separate disclosed observations from product marketing.
UK NCSC secure AI guidance
Government guidance for secure AI development across design, deployment and operation, with practical lifecycle considerations.
NIST AI Risk Management Framework
Official risk-management publications and version updates. Useful for governance planning, rather than as a product certification.
REST API endpoints for global security advisories
Official API documentation for retrieving and filtering global advisories, with affected-package data and source references.
Audit shows that safetensors is safe and ready to become the default
The project sponsor explains serialization risks and links the commissioned audit. Audit results apply to the reviewed scope and version.
Hugging Face Hub security documentation
Platform-maintainer documentation on model artifact scanning and repository security; useful when evaluating downloaded AI assets.
MITRE ATLAS tactics, techniques, and case studies data
The maintainers’ structured ATLAS data supports threat-model lookups and agent workflows. Preserve technique identifiers and linked case-study provenance.
GenAI and Agentic AI Exploit Roundup Q3 2026
An OWASP-hosted editorial roundup linking reported incidents and research. The page states it is neither exhaustive nor an official OWASP publication.
NVIDIA Product Security
NVIDIA’s official security bulletin repository provides maintainer notices and machine-readable records for its products, including AI infrastructure.
arXiv Cryptography and Security
Find recent security preprints and original papers. A posting is not peer-review approval; inspect methods, revisions and linked artifacts.
OWASP Cheat Sheet Series
OWASP application-security guides, including MCP, AI agent security and prompt injection. Each cheat sheet links practical controls and further references.
1Password's AI patching benchmark is misleading
A Trail of Bits post presenting a critical examination of 1Password's AI patching benchmark and arguing that it is misleading.
Auditing in the age of (good enough) AI
Trail of Bits describes using AI agents to build auditing tools and formal models for a Miden VM security review. Useful for exploring AI-assisted audit workflows.
Building an AI Detection Engine That Understands Agent Intent
Wiz describes its model-input/output telemetry pipeline for detecting changes in agent intent. This is the vendor's engineering account; SecLLM has not independently evaluated its effectiveness.
From Concept to Context Engine: How Wiz Built AI-Powered Data Discovery
Engineering article on the multi-agent pipeline and feedback loops Wiz used to turn a bucket scanner into an AI-powered data discovery context engine.
OWASP GenAI Security Project Unveils 2026 Top 10 for LLM Applications, New Agent Control Standard and Sponsors as Community Tops 30,000 Members
OWASP GenAI Security Project announces its 2026 Top 10 for LLM Applications, a new Agent Control Standard, and additional resources for securing generative and agentic AI.
The Infostealer Incursion: How Stolen Credentials Breach Cloud, Code, and AI Environments
Wiz Research analyzes NordStellar data to map credentials targeted by infostealer families and assess their potential impact across cloud, code, and AI environments.
Wiz Red Agent Finds Its Way Into Snowflake’s Internal Jira Through a Flaw in a GitHub Copilot–Assisted PR
Wiz reports a GitHub Actions injection flaw in Snowflake. Its correction says AI authorship of the vulnerable change is unconfirmed; Copilot Autofix made a separate fix in the same pull request.
No matching resources
Try another keyword or clear your filters.
Newest sorts by original publication date. Resources without a date appear last.
Directory updated
MCP ACCESS
Search these resources
from your agent.
Get source links, topic tags, and dates when available.