Safe labs

Practice the attack without running the attack.

Each lab uses a static simulated trace so developers can learn the pattern without exposing a real repo or executing untrusted code.

Install HOL GuardProtect a team

Safe labs

Practice attack patterns with static simulations. Nothing dangerous executes.

prompt-injectionStatic simulation

Prompt injection replay

See how hidden text tries to override your instructions.

Safe trace$ Agent reads issue text$ Hidden instruction appears$ Guard asks for reviewOpen the labmcp-tool-poisoningStatic simulation

MCP tool poisoning demo

Inspect a fake MCP tool description before it can mislead an agent.

Safe trace$ Tool advertises safe name$ Description asks for tokens$ Guard flags mismatchOpen the labsecret-exfiltrationStatic simulation

Secret exfiltration drill

Practice spotting a secret read request without exposing real secrets.

Safe trace$ Command requests env file$ Path is redacted$ Guard blocks before readOpen the labsupply-chainStatic simulation

Supply-chain install check

Walk through a simulated install that tries to change agent trust.

Safe trace$ Package install starts$ Postinstall wants shell access$ Guard pauses actionOpen the lab