Struct ShadowProbeExecutor 

pub struct ShadowProbeExecutor<T: ToolExecutor> { /* private fields */ }

Wraps an inner ToolExecutor and applies an LLM safety probe before high-risk calls.

ShadowProbeExecutor<T> is Clone when T: Clone (not required for operation). All methods delegate to inner after a probe verdict of Allow or Skip.

Concurrency

The probe field is Arc<dyn ProbeGate>, so multiple ShadowProbeExecutor instances sharing the same underlying ShadowSentinel (e.g., during parallel tool dispatch) are safe.

Implementations

impl<T: ToolExecutor> ShadowProbeExecutor<T>

pub fn new( inner: T, probe: Arc<dyn ProbeGate>, turn_number: Arc<AtomicU64>, risk_level: Arc<RwLock<String>>, ) -> Self

Create a new ShadowProbeExecutor wrapping inner.

Arguments

• inner — the next executor in the chain (typically PolicyGateExecutor).
• probe — the safety probe gate backed by ShadowSentinel.
• turn_number — shared atomic counter updated by the agent loop.
• risk_level — shared risk level string updated by the agent loop.

Trait Implementations

impl<T: ToolExecutor + Debug> Debug for ShadowProbeExecutor<T>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<T: ToolExecutor> ToolExecutor for ShadowProbeExecutor<T>

async fn execute(&self, response: &str) -> Result<Option<ToolOutput>, ToolError>

Legacy fenced-block path: probe not applied (no structured tool id).

async fn execute_confirmed( &self, response: &str, ) -> Result<Option<ToolOutput>, ToolError>

Legacy confirmed path: probe not applied.

async fn execute_tool_call( &self, call: &ToolCall, ) -> Result<Option<ToolOutput>, ToolError>

Structured tool call path: probe is applied before delegation.

Returns ToolError::SafetyDenied if the probe returns Deny. Delegates to inner on Allow or Skip.

async fn execute_tool_call_confirmed( &self, call: &ToolCall, ) -> Result<Option<ToolOutput>, ToolError>

Confirmed structured path: probe is still applied.

User confirmation does not bypass the safety probe — they are orthogonal gates.

fn tool_definitions(&self) -> Vec<ToolDef>

Return the tool definitions this executor can handle.

fn set_skill_env(&self, env: Option<HashMap<String, String>>)

Inject environment variables for the currently active skill. No-op by default.

fn set_effective_trust(&self, level: SkillTrustLevel)

Set the effective trust level for the currently active skill. No-op by default.

fn is_tool_retryable(&self, tool_id: &str) -> bool

Whether the executor can safely retry this tool call on a transient error.

fn is_tool_speculatable(&self, tool_id: &str) -> bool

Whether a tool call can be safely dispatched speculatively (before the LLM finishes).

fn requires_confirmation(&self, call: &ToolCall) -> bool

Return true when call would require user confirmation before execution.