LAB · Security
Prompt Injection Threat Lab
Classify trust boundaries, assets, actions, and mitigations for direct and indirect attacks.
- Suggested time
- 15 min
- Output artifact
- A minimal threat model with preventive, detective, and recovery controls.
- Privacy
- Scenarios remain local.
Local educational tool. Validate every result before production use.
When to use it
Threat lab
Use it before connecting untrusted content to tools, private data, or actions with external effects.
How to use it well
- 1
Enter a real case without secrets or personal data.
- 2
Run, inspect, and challenge the assumptions; the output is not automatic truth.
- 3
Save the artifact in your project and turn it into a reviewable test or decision.
Related foundations
Harness Engineering
08 — LLM System Security and Governance
Associated Lab: ../labs/07_prompt_injection.py
Open foundation20 minLab 07 — Prompt injection: measure attacks and apply layered controls
Localized walkthrough. Run the canonical public lab locally and preserve the result as evidence.
Open foundation3 min02 — Context, capabilities, and isolation
A harness does not improve by supplying more tokens. It improves when the agent can discover the right context or capability at the right moment, and when the system prevents that capability from exceeding the task's aut
Open foundation