Struct Proposal 

pub struct Proposal {
    pub approach: String,
    pub key_tradeoff: String,
    pub risks: Vec<String>,
    pub touches: Vec<String>,
    pub why_not_naive: String,
}

One advisor’s independent design proposal, gathered by crate::advise between a magi plan interview and the task file it files.

No patch and no plan of tool calls - a design proposal is cheap exactly because it is a few paragraphs an agent can write without touching the repository, which is what makes running three of them, on every task, affordable in a way three full implementations were not.

Fields

approach: String

What to do and how, in the advisor’s own words.

key_tradeoff: String

The one tradeoff this design turns on.

risks: Vec<String>

What could go wrong.

touches: Vec<String>

Files or modules the design touches.

why_not_naive: String

Why this earns its complexity over the obvious first draft - the question that keeps a proposal from being a restatement of the task.

Implementations

impl Proposal

pub fn validate(&self) -> Result<()>

Reject a proposal with an empty field that matters, so a malformed or lazy answer is retried instead of silently thinning the panel down to prose nobody can act on.

Trait Implementations

impl Clone for Proposal

fn clone(&self) -> Proposal

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Proposal

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Proposal

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Proposal

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

where __S: Serializer,

Serialize this value into the given Serde serializer.