How agentichardening/awesome-agentic-hardening works
This repository is a curated resource guide, not an executable application. A reader uses README.md to navigate from threat categories through hardening approaches and evaluation resources to governance standards. The README explicitly describes this organizing pipeline; the tree also contains taxonomy and resource-type documents, but their contents were not provided, so their internal structure is not represented.
Main entry points: README.md Curated Guide
Components 14
Threat Landscape
- Prompt Injection
README.md - Tool Misuse
README.md - Memory Poisoning
README.md - Protocol Threats
README.md - Identity & Supply
README.md
Hardening Techniques
- Prompt Hardening
README.md - Runtime Sandboxing
README.md - Detection & Monitoring
README.md - Protocol Hardening
README.md
Evaluation & Testing
- Red Teaming
README.md - Datasets & Research
README.md
Governance & Standards
- Standards & Compliance
README.md
Other components
- Reader
- Curated Guide
README.md
Connections 16
- Reader → Curated Guide: reads
README.md - Curated Guide → Prompt Injection: lists
README.md - Curated Guide → Tool Misuse: lists
README.md - Curated Guide → Memory Poisoning: lists
README.md - Curated Guide → Protocol Threats: lists
README.md - Curated Guide → Identity & Supply: lists
README.md - Curated Guide → Prompt Hardening: lists
README.md - Curated Guide → Runtime Sandboxing: lists
README.md - Curated Guide → Detection & Monitoring: lists
README.md - Curated Guide → Protocol Hardening: lists
README.md - Curated Guide → Red Teaming: lists
README.md - Curated Guide → Datasets & Research: lists
README.md - Curated Guide → Standards & Compliance: lists
README.md - Prompt Injection → Prompt Hardening: informs
README.md - Prompt Hardening → Red Teaming: evaluated by
README.md - Red Teaming → Standards & Compliance: leads to
README.md