pub struct EffortSchedule {Show 14 fields
pub code: Option<String>,
pub round_1: Option<String>,
pub rest: Option<String>,
pub triage: Option<String>,
pub implement: Option<String>,
pub review_1: Option<String>,
pub review_rest: Option<String>,
pub respond: Option<String>,
pub close: Option<String>,
pub screen: Option<String>,
pub checkin: Option<String>,
pub split: Option<String>,
pub brainstorm: Option<String>,
pub recheck: Option<String>,
}Fields§
§code: Option<String>Drafting an issue for spar code.
round_1: Option<String>The deep first review.
Kept, and still read by every call that has no key of its own, so an existing config keeps working exactly as it did.
rest: Option<String>Every round after the first, and the closing pass. Both are asked a narrower question than the first review, so neither buys its depth again.
triage: Option<String>Both agents, over the whole queue. The most expensive call in a run, and the one the two key schedule never mentioned.
implement: Option<String>§review_1: Option<String>§review_rest: Option<String>§respond: Option<String>§close: Option<String>§screen: Option<String>§checkin: Option<String>§split: Option<String>§brainstorm: Option<String>Every brainstorm round. Diverging is the deep call and the later rounds answer narrower questions, but they read the same key: a brainstorm is cheap next to a run, and one word for it is enough.
recheck: Option<String>A second look at a reviewed pull request: a delta and a list of claims
to check, so it sits with the later review rounds and reads
review_rest when it has no word of its own.