Skip to content

Security and Safety

4 build exercises to practise the concepts in this domain.

7.1 — Draw the Trust Boundary, Then Close It in Four Passes

Advanced
45 minutes
  • How to inventory every channel by which untrusted text reaches a context window
  • How to deliver third-party content as a labelled, JSON-encoded tool_result
  • How to cut a tool surface by removal, scoping and credential narrowing
  • How to add a screening call with structured outputs and branch on the parsed verdict
  • How to treat your own logs and your provider arrangement as part of the PII decision

7.2 — Name the Enforcement Point for Every Layer You Claim to Have

Intermediate
35 minutes
  • How to audit a guardrail set by enforcement point rather than by intention
  • How to decide fail-closed behaviour for each consequential action in writing
  • How to narrow a credential and prove the narrowing
  • How to check a deployment against the high-risk obligations in the Usage Policy

7.3 — One Rule, One Hook, One Proof That It Fired

Beginner
25 minutes
  • How to tell a hook-shaped rule from a preference
  • How to register a PreToolUse hook with a matcher narrow enough to be safe
  • How the exit-code contract decides whether a gate blocks or quietly fails open
  • How to name what a given hook will never catch

7.4 — Give Every Caller Its Own Identity, Then Prove You Could Investigate One

Intermediate
35 minutes
  • How to inventory every place a Claude credential currently exists
  • How to split a shared key into per-service, per-workspace identities
  • How to run the exposure playbook in the documented order against a disposable key
  • How to turn on the attribution you would need during an incident