Struct QueryContext 

pub struct QueryContext {
    pub tool_name: String,
    pub args: Vec<String>,
    pub tool_input: Value,
    pub hook_type: Option<String>,
    pub agent_name: Option<String>,
    pub fs_op: Option<String>,
    pub fs_path: Option<String>,
    pub net_domain: Option<String>,
    pub mode: Option<String>,
}

The context passed to the evaluator, extracted from a tool invocation.

Fields

tool_name: String

The tool name (e.g. “Bash”, “Read”).

args: Vec<String>

Positional args (for Bash: parsed command parts).

tool_input: Value

The full tool_input JSON.

hook_type: Option<String>

Hook type, if this is a hook invocation.

agent_name: Option<String>

Agent name, if this is an agent invocation.

fs_op: Option<String>

Capability: filesystem operation (“read” or “write”), if applicable.

fs_path: Option<String>

Capability: resolved filesystem path, if applicable.

net_domain: Option<String>

Capability: network domain, if applicable.

mode: Option<String>

Permission mode from Claude Code (e.g. “default”, “plan”).

Implementations

impl QueryContext

pub fn from_tool(tool_name: &str, tool_input: &Value) -> Self

Build a QueryContext from a tool invocation.

Trait Implementations

impl Debug for QueryContext

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

Formats the value using the given formatter.