Skip to main content

Flagship 01 / Sley

Software change, made structural.

Sley gives autonomous software a structural view of a program before modification, then keeps intent, authority, and verification visible to human reviewers.

Sley structural graph flowing through typed language layers
SourceStructureAuthorityVerification
examples/project/src/app/main.sleyACTUAL SOURCE
module app.compute

task classify -> Text {
  take score: Int
  return if score >= 90 { "excellent" } else { "steady" }
}

task main -> Text {
  bind score = 40 + 55
  return call classify(score)
}

Source is the entry point

Readable syntax. Structural reports. Separate authority.

This example is real Sley source from the public repository. Its call structure can be queried before any write is authorized.

Why it exists

Text is not the whole program.

Conventional editing asks a system to infer structure from text, change it, and explain the result afterward. Sley makes structure part of the language surface before mutation begins.

Terminal evidence

sley query --json --kind calls examples/project

"schema": "sley.query.report.v0"

"kind": "calls"

"entry_module": "app.main"

"modules": 2, "calls": 1

A bounded change lifecycle
  1. 01Inspect structure
  2. 02Plan transformation
  3. 03Authorize scope
  4. 04Verify evidence
local v1 checks
23 / 23
report schemas
42
contract fixtures
128
corpus cases
66
integration checks
199

Public checkpoint recorded August 22, 2026. Later worktree state may differ.

What systems can inspect

  • Typed structural nodes and relationships
  • Declared intent and transformation boundaries
  • Available capabilities before a write
  • Deterministic reports and verification results

What people retain

  • Readable source and an explicit proposed change
  • Authority over whether the proposal may execute
  • A reviewable boundary between reports and mutation
  • Evidence tied to the exact checkpoint being discussed