Struct ToolResult 

pub struct ToolResult {
    pub kind: String,
    pub command_id: String,
    pub run_stream: StreamRef,
    pub call_id: String,
    pub text: String,
    pub correlation_facts: Option<Value>,
    pub edit_facts: Option<Value>,
    pub extra: Map<String, Value>,
}

tool.result — outcome of one tool invocation (live providers only).

No task_id, but the wire models each tool call as its own task (task_kind: tool.<tool_name>) and correlation_facts.tool_name names it — match on that, latest-first. Recency-of-running-tasks heuristics mis-attribute: the issuing tool task has already completed when this record lands. call_id is the provider’s call id, not a task handle — see the README’s known-wire-gaps section.

correlation_facts is absent on some tool results (e.g. compact bash results like {"items":5,"ok":true,"revision":4} observed on 3035c77c-efca...).

text is opaque for most tools (prose, e.g. write_file → "wrote 6 bytes …"), but the bash/command tool packs a structured JSON object into text (see CommandResult). Use ToolResult::command_result to get a typed view when that shape is present. The same JSON is also emitted as a task.lifecycle.output chunk — consumers that render both channels should de-dupe (the tool.result record is authoritative).

Fields

kind: String

command_id: String

run_stream: StreamRef

call_id: String

Provider call id this result answers.

text: String

Result text as shown to the model (including failure prose). For the bash/command tool this is a JSON string of a CommandResult; see ToolResult::command_result and ToolResult::try_command_result.

correlation_facts: Option<Value>

Correlation summary — observed {outcome, tool_name}, open-shaped. Absent on some results; treat as None when missing.

edit_facts: Option<Value>

Populated for file-editing tools; open-shaped.

extra: Map<String, Value>

Implementations

impl ToolResult

pub fn is_command_tool(&self) -> bool

Whether this result came from the bash/command tool, via correlation_facts.tool_name == "bash" | "command".

pub fn command_result(&self) -> Option<CommandResult>

Try to parse text as a structured CommandResult (the bash tool shape described in #294). Returns None if text is not valid JSON for that shape. Checks is_command_tool() first but also accepts any JSON object that deserializes as CommandResult — so compact results without correlation_facts (observed on #299) still parse.

pub fn try_command_result(&self) -> Result<CommandResult, Error>

Fallible parse of text as CommandResult, preserving the serde error.

Trait Implementations

impl Clone for ToolResult

fn clone(&self) -> ToolResult

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ToolResult

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ToolResult

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ToolResult

fn eq(&self, other: &ToolResult) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for ToolResult

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ToolResult