Repository navigation
llm-shield-proxy 1.7.0
llm-shield-proxy 1.7.0. Minor release: request-redaction and credential-routing fixes, with behaviour changes an operator should know about.
What is in it since 1.6.11
-
#208 Every value inside a chat message is scanned, whatever its shape. Real shapes that reached the provider unredacted:
- OpenAI custom tool calls (
custom.input); - Anthropic
server_tool_useandmcp_tool_useinput; - and any value of an unexpected type under a key the walk owns.
Gemini's
thought_signature, media payloads and file ids still go out unchanged. - OpenAI custom tool calls (
-
- Client-picked hosts. A host the caller picks with
X-Upstream-Base-UrlorX-Shield-Fallback-Urlgets a proxy-held key only if it is the configured upstream host or a known provider's API host, and then only that provider's key. A failover carries only the fallback's key. - The canary tripwire fires.
CANARY_TOKENnow goes into the prompt, and it is matched in streamed text too. - The raw-text reply fallback runs the response PII scan.
GETpass-through survives a compressed reply.- An Anthropic caller's own
x-api-keyis no longer blanked.
- Client-picked hosts. A host the caller picks with
-
#196 (merged 2026-10-07, first release): the provider's error reason is relayed to the client, scrubbed.
-
#210 README and docs: LiteLLM 1.106.0 and later ship the built-in guardrail.
Behaviour changes
- With
ALLOW_CLIENT_UPSTREAM_OVERRIDE=true, a virtual-key orOVERRIDE_CLIENT_AUTHrequest naming any other host now gets a 403. A BYOK caller's own key is unaffected. - With
ENABLE_CANARY_TRIPWIRE=true, the system prompt now carriesCANARY_TOKEN.
pii-leak-benchmark is unchanged at 0.6.0.