II.
ContentPolicy overview
Reference · livecontent-policy:default-acceptable-use
Default acceptable-use policy overview
Inspect the raw attributes, linked wiki pages, and inbound or outbound graph edges for content-policy:default-acceptable-use.
Attributes
displayName
Default acceptable-use policy
policyKind
deny-list
severity
blocking
rules
Block any output that:
- depicts CSAM (zero tolerance);
- provides operational uplift for weapons of mass destruction;
- constitutes targeted harassment of an identifiable person.
notes
Platform-wide content rules applied to every agent run.
Outgoing edges
applied_by1
- agentVersion:claude:ge-0-0-0·AgentVersionClaude Code
detects_jailbreak4
- jailbreak:dan·JailbreakPatternDAN (Do Anything Now)
- jailbreak:prompt-injection-indirect·JailbreakPatternIndirect prompt injection via retrieved content
- jailbreak:many-shot·JailbreakPatternMany-shot jailbreak
- jailbreak:base64-obfuscation·JailbreakPatternBase64 / encoded-payload obfuscation
Incoming edges
applies_policy1
- agentVersion:claude:ge-0-0-0·AgentVersionClaude Code
assesses_policy4
- tool-descriptor:babysitter-sdk-process-api·ToolDescriptorBabysitter SDK Process API
- tool-descriptor:bash·ToolDescriptorBash
- tool-descriptor:monitor·ToolDescriptorMonitor
- tool-descriptor:powershell·ToolDescriptorPowerShell
evaluates_policy1
- benchmark:harmbench·BenchmarkHarmBench