Struct AgentRun 

pub struct AgentRun {
    pub agent: String,
    pub model: String,
    pub run_id: String,
    pub started_at: String,
    pub finished_at: Option<String>,
    pub context: BTreeMap<String, String>,
    pub tool_calls: Vec<ToolCallTrace>,
    pub permissions: Option<PermissionState>,
}

Agent provenance attached to a StateProposal.

Doctrine: the substrate stays model-agnostic. Agents — Literature Scout, Notes Compiler, Code Analyst, etc. — sit in the vela-scientist crate (or external code) and write proposals into a frontier through the existing protocol. This struct is the reviewer-facing record of who proposed what, with what model, during which run — never used as access control or trust assignment.

Fields

agent: String

Stable agent name (e.g. “literature-scout”). Pairs with the proposal’s actor.id == "agent:literature-scout".

model: String

Model identifier (e.g. “claude-sonnet-4-6”). Free-form so the substrate never has to enumerate model names.

run_id: String

Run identifier — typically a UUID or short hash. Lets the reviewer group multiple proposals that came out of the same agent invocation.

started_at: String

ISO-8601 wall-clock start of the run.

finished_at: Option<String>

ISO-8601 wall-clock end. Optional because some agents stream.

context: BTreeMap<String, String>

Free-form context the reviewer should see — e.g. the input folder path, the count of papers processed, the prompt version. Kept as a flat string map so it round-trips cleanly through canonical JSON without imposing a schema.

tool_calls: Vec<ToolCallTrace>

v0.49: explicit tool-call traces from this run. Each entry records one tool invocation by content-addressable summary (tool name + input hash + output hash + duration). Lets a reviewer see what the agent actually called without bloating the bundle with raw payloads. Optional + skip-if-empty so existing frontiers round-trip byte-identically.

permissions: Option<PermissionState>

v0.49: declared permission state for this run. Lists the data sources the agent had read access to and the tools it could invoke. Reviewers compare this declaration against tool_calls to spot drift. Optional + skip-if-empty so existing frontiers round-trip byte-identically.

Trait Implementations

impl Clone for AgentRun

fn clone(&self) -> AgentRun

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AgentRun

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

Formats the value using the given formatter.

impl Default for AgentRun

fn default() -> AgentRun

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for AgentRun

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for AgentRun

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for AgentRun

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for AgentRun

impl StructuralPartialEq for AgentRun