PCD agent-readable summary

PCD describes selected software logic as a bounded circuit with explicit inputs, outputs, branches, operations, and fallback behavior.

Documentation

Describe the logic before reviewing the code.

PCD turns a bounded requirement into an inspectable circuit of inputs, decisions, operations, outputs, and fallback behavior.

approval_boundary.pcd

input amount: Money
input role: Role

branch role == guest
  return amount >= 1000

fallback deny
role
compare
approve
deny

A small language for a bounded decision.

PCD keeps the review unit and its domain boundaries narrow enough to understand and explicit enough to reproduce.

01

Inputs and outputs

Name what enters the selected circuit and what it must return.

02

Monomers

Select supported atomic operations instead of hiding behavior inside ad hoc code.

03

EVA composition

Keep sequence, branches, and parallel paths visible as structure.

04

Fail closed

Declare the fallback when an input or assumption falls outside the reviewed scope.

From requirement to reviewable topology.

The same bounded description can support local authoring, topology inspection, and a team review handoff.

01

Bound

Write the decision, inputs, outputs, and assumptions.

02

Compose

Choose monomers and make branch structure explicit.

03

Inspect

Use the CLI to read the PCD and retain command evidence.

04

Review

Carry the selected blueprint into Platform when collaboration is required.

PCD makes declared logic inspectable. It does not by itself certify the generated implementation or the whole application.

Make one software decision explicit.

Start with a bounded requirement, then use the public language docs and CLI to create a reviewable blueprint.