REFLEX Battlecards
Each battlecard takes one real attack pattern and walks it through the six REFLEX stages. You get a concrete scenario, the real incident behind it, and what to do about it at each stage.
They are grouped by where the attack lands. Start with the area you own.
Package Ecosystems
Attacks on the registries and packages you install: typosquats, hijacked maintainers, dependency confusion and malicious install scripts.
- Abandoned Package Hijack
- chalk and debug npm Compromise (2025)
- Dependency Confusion Attacks
- End-of-Life Dependencies
- Go Modules Dependency Risks
- Industrialised Malware Campaigns
- Install Scripts and Archive Extraction Tricks
- Lockfile and SemVer Range Abuse
- Maintainer Account Takeover
- Malicious Third-Party Mobile SDKs
- Node.js and npm Dependency Sprawl
- Python and pip Supply Chain Attacks
- Registry Namespace Collisions
- Rust Crates Supply Chain
- Shai-Hulud npm Worm
CI/CD & Build
Pipelines, runners, caches and build tooling: the places where your code turns into something you ship.
- Build Cache and Artifact Poisoning
- Build Tooling Blind Spots
- Cloud Metadata Credential Theft in Builds
- Compromised Security Scanners and Cascading Credential Theft
- curl | bash Installer Attacks
- Docker Socket and Privileged Builds
- Environment Variable and Runtime Hook Hijacking
- GitHub Actions Workflow Poisoning
- GitLab CI Runner Attacks
- Hosted CI Secret Theft (CircleCI, Travis CI)
- Jenkins CI/CD Compromise
- Maven and Gradle Wrapper Poisoning
Source Control & Review
Branch protection, code review and the code itself: where a malicious change has to get past people.
- Branch Protection and CODEOWNERS Gaps
- Homoglyph and Trojan Source Attacks
- Submodule and Vendored Code Drift
Developer Environment
Your laptop, IDE and dev containers: the first place an attacker can run code as you.
- Agent Instruction and Config File Poisoning
- Browser Extension Publisher Takeover
- Dev Container Lifecycle Hooks
- Malicious IDE Extensions
Secrets & Credentials
Keys, tokens and SSH agents, and how they leak.
- Exposed Secrets in Code and Config
- SSH Agent Forwarding and Host Key Trust
- Third-Party Integration Token Theft
Infrastructure & Cloud
Cloud identity, Kubernetes, images and infrastructure as code.
- Container Base Image and Registry Risks
- Exposed Developer Consoles and Debug Endpoints
- GitHub OIDC Misconfiguration
- Kubernetes Misconfigurations
- Packer Build Script Risks
- Terraform Module Attacks
Network & Distribution
How dependencies and scripts reach you, and who can tamper with them on the way.
- CDN Script Compromise and Mirror Substitution
- DNS Hijacking and Man-in-the-Middle on Dependency Downloads
- Hijacked Auto-Update Channels
- Shadow Downloads
AI/ML Security
Models, datasets, AI services and the agents now working in your pipelines.
- Agent Tool and MCP Server Abuse
- AI Agents and Models in CI/CD Pipelines
- AI Dependency Typosquatting and Slopsquatting
- AI-Recommended Dependencies and Hallucinated Versions
- Data Poisoning Attacks
- Dataset Supply Chain Attacks
- Leaked AI API Keys and Cost Amplification
- LLMjacking with Stolen Cloud Credentials
- Model Provenance and Evaluation Bypass
- Model Supply Chain Metadata Leaks
- Poisoned AI/ML Models
- Prompt Injection in AI Systems
- Vulnerable AI Framework Code
Supply Chain Integrity
Provenance, SBOMs and licences: knowing what you ship and whether you can trust it.
- Code-Signing Certificate Theft and Misissuance
- Licence Changes and Protestware
- Provenance and SBOM Tampering
- Vulnerability Data Gaps
Social Engineering
Attacks on the people: phishing, fake hires and the slow cultivation of trust.
- Developer Review Fatigue and Social Engineering
- Fake Remote Developers and Insider Access
- Phishing Developers
67 battlecards. Each one follows the REFLEX framework: Reconnaissance, Evaluate, Fortify, Limit, Expose, Exercise.