Struct ExecuteRequest 

pub struct ExecuteRequest {
    pub model: Option<String>,
    pub provider: Option<String>,
    pub model_ref: Option<ProviderModelRef>,
    pub skill_mode: Option<String>,
    pub reasoning_effort: Option<ReasoningEffort>,
    pub client_sync: Option<ExecuteClientSync>,
    pub no_human_approver: bool,
}

Request payload for agent execution.

Fields

• model - Optional model identifier used only as a compatibility fallback

Note

The model parameter is optional.

Runtime behavior is session-driven:

• prefer session.model
• then provider default model
• finally request.model as a compatibility fallback

Examples

{
  "model": "claude-3-opus"
}

Fields

model: Option<String>

Optional model compatibility fallback for execution.

provider: Option<String>

Optional provider name (used with provider_model_ref feature).

model_ref: Option<ProviderModelRef>

Optional provider+model reference (takes priority when present).

skill_mode: Option<String>

Optional per-execution skill mode override (for example: “code”, “ask”).

When provided, skill discovery prefers skills-<mode> directories.

reasoning_effort: Option<ReasoningEffort>

Optional reasoning effort override for this execution.

When omitted, the active provider default from config is used.

client_sync: Option<ExecuteClientSync>

Optional server-confirmed client cursor used for pre-execution sync checks.

no_human_approver: bool

Whether this run has NO interactive human approver (headless -p, a scheduled job, a deployed broker). #74: this is re-derived per user-initiated execute and OVERWRITES the session’s persisted no_human_approver, so a session first run headlessly and later reopened interactively (UI omits this → false) correctly resets to the human-present posture. Suspend/resume does NOT go through this handler, so a within-run resume keeps the persisted posture.

Trait Implementations

impl<'de> Deserialize<'de> for ExecuteRequest

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.