PowerShell Completion Predictor

Why it's ranked here

The design is practical and unusually instructive: it reuses native completion behavior while isolating prediction work from the interactive session. State synchronization makes results relevant to the user’s environment, and Git-specific logic adds real value. Its experimental status and deliberately narrow prediction coverage keep the verdict mixed.

What's good

A busy prediction session drops new work instead of allowing unsafe reentrancy. The tool filters unchanged results, preserves distinct tooltips, handles local path prefixes, and caps output volume. It also notices accepted command lines so cached repository information can be refreshed. The code doubles as a compact example for predictor authors.

Tradeoffs

Command-name prediction is intentionally disabled because discovery commonly exceeds the response deadline. Broad argument prediction is also constrained because filesystem enumeration and external custom completers can be slow. Git option prediction remains limited, concurrent keystroke requests may be ignored while the prediction session is occupied, and the project explicitly remains experimental.

How to use it well

Use it if you work interactively in modern PowerShell and want native completion candidates surfaced earlier, especially around paths and supported Git operations. It suits users comfortable installing a module and enabling plugin predictions. Do not treat it as comprehensive command discovery, broad Git flag assistance, or a stable finished product.

Observed

License

MIT License

Primary language

C#

Interface

PowerShell command-line prediction plugin registered through the subsystem plugin model

Installation

Distributed as a PowerShell module installable from PowerShell Gallery

Runtime requirements

PowerShell 7.2 or above and PSReadLine 2.2.2 or above

Build surface

.NET 6 SDK with a PowerShell build script

Packaging

PowerShell module manifest with a nested compiled assembly