Shrt

Shrt is a contract-driven gRPC regression testing tool for AI coding agents, with human-approved baselines.

Shrt is a regression testing tool designed to test backends through chains of API calls defined in YAML files. It is built to be driven by coding agents, such as through an installed Claude Code skill, while supporting manual command line usage. The tool speaks the Connect protocol, using JSON over HTTP POST, and supports unary and server-streaming RPCs by compiling proto files with buf. It features commands to initialize configurations, generate contracts and chains, run tests, and approve passing runs as safe baselines for future releases. Shrt targets local environments by default to prevent accidental data modification in production systems. It requires Go 1.26 or later and buf on the PATH to compile backend proto files.

Shrt at a glance

Best for: Software engineer
Runs on: CLI, self-hosted
Pricing: Free
Released: September 2026
Category: AI coding tool
Website: github.com/n4darae/shrt

Key features

Compiles backend proto files with buf
Supports Connect protocol over HTTP POST
Generates API test chains from contracts
Establishes human-approved response baselines
Detects regression fields and suspects
Installs a Claude Code skill for agent use
Blocks requests to non-localhost addresses

Use it for

Testing backend APIs through chains of calls
Creating test chains automatically from contracts
Approving passing test runs as safe baselines
Verifying releases against established baselines

Alternatives to Shrt

jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
pstack — Improves AI coding agent outcomes with rigorous workflows.
ZCode — Integrates GLM-5.3 with multi-agent workflows and remote chat control.
TestSprite — Runs end-to-end tests on live apps to catch regressions without manual scripting.
Shim — It offers zero runtime dependencies, no telemetry, and a static binary.
Parasoft — AI-driven test automation and static analysis for software quality at scale.
Gspot — Stops bad code, type errors, and leaked keys before commits happen.
abtop — Monitor multiple AI agent sessions and detect orphan ports from a single screen.
Pawl — Stops costly agent mistakes using fast deterministic code instead of prompt instructions.
Eval Skills — Guides coding agents from first execution to validated graders and experiments.
Weft — It enables safe parallel coding agent work without file locks.
Ctxfw — Reduces AI context token mass while preserving code structure and dependencies.

All Shrt alternatives