Struct TestCase 

pub struct TestCase {
    pub name: String,
    pub skill: PathBuf,
    pub input: String,
    pub user: Option<SimulatedUser>,
    pub mocks: Vec<MockDecl>,
    pub spy: bool,
    pub evals: Vec<Eval>,
}

One test case.

This type is the source of truth for the input contract (schemas/case.schema.json): the shape a --case-json payload — and the SDKs’ generated case models — must have. Serialization skips absent/default fields so the canonical JSON form is minimal; the SDK case builders emit that same form (pinned by the kitchen-sink golden in tests/fixtures/contract/).

Fields

name: String

Human-readable name (defaults to the file stem when loaded from a file).

skill: PathBuf

Path to the skill directory under test, relative to the test-case file.

input: String

The initial data/prompt handed to the skill as the first user message.

user: Option<SimulatedUser>

Present for multi-turn cases; absent for single-turn.

mocks: Vec<MockDecl>

Mock/spy declarations for this case: a declaration with a stub/deny/ rewrite action intercepts matching tool calls; one without observes only. called/not_called evals reference these by name.

spy: bool

Record every tool call through the mock/spy channel even with no mocks declared, so code-level consumers (the SDKs’ spies) get records. Implied whenever mocks is non-empty.

evals: Vec<Eval>

The evals that decide whether this case passes. Must be non-empty.

Implementations

impl TestCase

pub fn load(path: &Path) -> Result<Self>

Load a test case from a YAML file. The name defaults to the file stem and skill is resolved relative to the file’s directory.

Errors

Error::Io if the file cannot be read, Error::Yaml on parse failure, and Error::Invalid if the case is internally inconsistent.

pub fn from_json(json: &str, base_dir: &Path) -> Result<Vec<Self>>

Parse one or more fully-specified test cases from JSON — a single case object or an array — the shape the language SDKs emit when a case is built in code rather than written as YAML (skilltest run --case-json).

Unlike load these cases have no source file, so each is finalized against base_dir (typically the working directory): a relative skill resolves there, and an unnamed case defaults to case (suffixed with its index when several are unnamed, keeping report keys distinct).

Errors

Error::Invalid if the JSON does not parse into test cases, or if any case is internally inconsistent.

pub fn finalize(&mut self, base_dir: &Path, default_name: &str) -> Result<()>

Finalize an in-memory case: default name to default_name when unset, resolve a relative skill against base_dir, and validate. Shared by load (anchored at the file’s directory) and from_json (anchored at the working directory, since a code-defined case has no file).

Errors

Error::Invalid if the case is internally inconsistent.

pub fn is_multi_turn(&self) -> bool

Whether this is a multi-turn case (has a simulated user).

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

Validate the case’s structure and every eval.

Errors

Error::Invalid when input/evals are empty or an eval is malformed.

Trait Implementations

impl Clone for TestCase

fn clone(&self) -> TestCase

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TestCase

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TestCase

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl JsonSchema for TestCase

fn schema_name() -> Cow<'static, str>

The name of the generated JSON Schema.

fn schema_id() -> Cow<'static, str>

Returns a string that uniquely identifies the schema produced by this type.

fn json_schema(generator: &mut SchemaGenerator) -> Schema

Generates a JSON Schema for this type.

fn inline_schema() -> bool

Whether JSON Schemas generated for this type should be included directly in parent schemas, rather than being re-used where possible using the $ref keyword.

impl PartialEq for TestCase

fn eq(&self, other: &TestCase) -> 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 TestCase

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 TestCase