examiner

Documentation

Overview

Package examiner implements the blind test examiner of §6.8. The examiner sees only the task contract and the public boundary (expone signatures, endpoints, CLI). It never sees function bodies.

Index

• Constants
• func Run(ctx context.Context, roomPath string, o Options) (bool, error)
• type Options
• type Runner
  • func (r Runner) Run(ctx context.Context, roomPath string) (bool, error)

Constants

const (
	VisibleFileName = "devclean_visible_test.go"
	HiddenFileName  = "devclean_hidden_test.go"
	DefaultTimeout  = 3 * time.Minute
)

Functions

func Run

func Run(ctx context.Context, roomPath string, o Options) (bool, error)

Run invokes the examiner agent, parses the response, writes the visible suite to the worktree and seals the hidden suite. Returns (true, nil) when a hidden suite was sealed. Returns (false, nil) on graceful degradation (no sealed suite written). Never returns an error that should stop the implementer.

Types

type Options

type Options struct {
	Agent   loop.Agent
	Task    task.Task
	Root    string
	Model   string
	Timeout time.Duration
}

Options configures one examiner run.

type Runner

type Runner struct{ Options }

Runner implements loop.Examinador so cmd/run can wire it without a cycle.

func (Runner) Run

func (r Runner) Run(ctx context.Context, roomPath string) (bool, error)