Graphorin API reference v0.12.1

---

Graphorin API reference / @graphorin/tools / / defaultInboundSanitization

Function: defaultInboundSanitization()

function defaultInboundSanitization(trustClass): InboundSanitizationPolicy;

Defined in: packages/tools/src/builder/trust-class.ts:66

Default inbound-sanitization policy per trust class. Operator overrides via tool({ inboundSanitization: ... }) always win.

• 'first-party-built-in' → 'pass-through' (zero overhead).
• 'first-party-user-defined' → 'detect-and-flag' (no body modification, but operator visibility is preserved).
• 'skill-trusted' → 'detect-and-flag'.
• 'skill-untrusted' → 'detect-and-strip-and-wrap'.
• 'mcp-derived' → 'detect-and-strip-and-wrap'.
• 'web-search' → 'detect-and-strip-and-wrap'.
• 'channel-inbound' → 'detect-and-strip-and-wrap' (message-borne channel text; no tool ever registers with this class, but the gateway reuses the same policy table).

Parameters

Parameter	Type
trustClass	ToolTrustClass

Returns

InboundSanitizationPolicy

Stable