pub struct AdversarialPolicyGateExecutor<T: ToolExecutor> { /* private fields */ }Expand description
Wraps an inner ToolExecutor, running an LLM-based adversarial policy check
before delegating structured tool calls.
Only execute_tool_call and execute_tool_call_confirmed are intercepted.
Legacy execute / execute_confirmed bypass the check (no structured tool_id).
Implementations§
Source§impl<T: ToolExecutor> AdversarialPolicyGateExecutor<T>
impl<T: ToolExecutor> AdversarialPolicyGateExecutor<T>
Sourcepub fn new(
inner: T,
validator: Arc<PolicyValidator>,
llm: Arc<dyn PolicyLlmClient>,
) -> Self
pub fn new( inner: T, validator: Arc<PolicyValidator>, llm: Arc<dyn PolicyLlmClient>, ) -> Self
Create a new AdversarialPolicyGateExecutor.
Sourcepub fn with_audit(self, audit: Arc<AuditLogger>) -> Self
pub fn with_audit(self, audit: Arc<AuditLogger>) -> Self
Attach an audit logger.
Trait Implementations§
Source§impl<T: ToolExecutor + Debug> Debug for AdversarialPolicyGateExecutor<T>
impl<T: ToolExecutor + Debug> Debug for AdversarialPolicyGateExecutor<T>
Source§impl<T: ToolExecutor> ToolExecutor for AdversarialPolicyGateExecutor<T>
impl<T: ToolExecutor> ToolExecutor for AdversarialPolicyGateExecutor<T>
Source§async fn execute(&self, response: &str) -> Result<Option<ToolOutput>, ToolError>
async fn execute(&self, response: &str) -> Result<Option<ToolOutput>, ToolError>
Parse
response for fenced tool blocks and execute them. Read moreSource§async fn execute_confirmed(
&self,
response: &str,
) -> Result<Option<ToolOutput>, ToolError>
async fn execute_confirmed( &self, response: &str, ) -> Result<Option<ToolOutput>, ToolError>
Execute bypassing confirmation checks (called after user approves). Read more
Source§fn tool_definitions(&self) -> Vec<ToolDef>
fn tool_definitions(&self) -> Vec<ToolDef>
Return the tool definitions this executor can handle. Read more
Source§async fn execute_tool_call(
&self,
call: &ToolCall,
) -> Result<Option<ToolOutput>, ToolError>
async fn execute_tool_call( &self, call: &ToolCall, ) -> Result<Option<ToolOutput>, ToolError>
Source§async fn execute_tool_call_confirmed(
&self,
call: &ToolCall,
) -> Result<Option<ToolOutput>, ToolError>
async fn execute_tool_call_confirmed( &self, call: &ToolCall, ) -> Result<Option<ToolOutput>, ToolError>
Execute a structured tool call bypassing confirmation checks. Read more
Source§fn set_skill_env(&self, env: Option<HashMap<String, String>>)
fn set_skill_env(&self, env: Option<HashMap<String, String>>)
Inject environment variables for the currently active skill. No-op by default. Read more
Source§fn set_effective_trust(&self, level: SkillTrustLevel)
fn set_effective_trust(&self, level: SkillTrustLevel)
Set the effective trust level for the currently active skill. No-op by default. Read more
Auto Trait Implementations§
impl<T> Freeze for AdversarialPolicyGateExecutor<T>where
T: Freeze,
impl<T> !RefUnwindSafe for AdversarialPolicyGateExecutor<T>
impl<T> Send for AdversarialPolicyGateExecutor<T>
impl<T> Sync for AdversarialPolicyGateExecutor<T>
impl<T> Unpin for AdversarialPolicyGateExecutor<T>where
T: Unpin,
impl<T> UnsafeUnpin for AdversarialPolicyGateExecutor<T>where
T: UnsafeUnpin,
impl<T> !UnwindSafe for AdversarialPolicyGateExecutor<T>
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more
Source§impl<T> ErasedToolExecutor for Twhere
T: ToolExecutor,
impl<T> ErasedToolExecutor for Twhere
T: ToolExecutor,
fn execute_erased<'a>( &'a self, response: &'a str, ) -> Pin<Box<dyn Future<Output = Result<Option<ToolOutput>, ToolError>> + Send + 'a>>
fn execute_confirmed_erased<'a>( &'a self, response: &'a str, ) -> Pin<Box<dyn Future<Output = Result<Option<ToolOutput>, ToolError>> + Send + 'a>>
fn tool_definitions_erased(&self) -> Vec<ToolDef>
fn execute_tool_call_erased<'a>( &'a self, call: &'a ToolCall, ) -> Pin<Box<dyn Future<Output = Result<Option<ToolOutput>, ToolError>> + Send + 'a>>
fn execute_tool_call_confirmed_erased<'a>( &'a self, call: &'a ToolCall, ) -> Pin<Box<dyn Future<Output = Result<Option<ToolOutput>, ToolError>> + Send + 'a>>
Source§fn set_skill_env(&self, env: Option<HashMap<String, String>>)
fn set_skill_env(&self, env: Option<HashMap<String, String>>)
Inject environment variables for the currently active skill. No-op by default.
Source§fn set_effective_trust(&self, level: SkillTrustLevel)
fn set_effective_trust(&self, level: SkillTrustLevel)
Set the effective trust level for the currently active skill. No-op by default.
Source§fn is_tool_retryable_erased(&self, tool_id: &str) -> bool
fn is_tool_retryable_erased(&self, tool_id: &str) -> bool
Whether the executor can safely retry this tool call on a transient error.