alfranz

alfranz / hush

Public

Context-efficient command runner for coding agents.

alfranz.github.iohush ai-agents cli coding-assistant developer-tools devops
13
0
100% credibility
Found Mar 02, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

Hush is a tool that executes shell commands and outputs a concise success marker or filtered failure details to minimize verbose logs, particularly useful for AI-assisted coding workflows.

How It Works

1
📰 Discover hush

You hear about hush, a handy tool that runs your project checks cleanly, showing just a checkmark when things go well or key details when they don't.

2
📥 Get hush set up

You easily download hush and add it to your computer's tools so it's ready to use anytime.

3
🚀 Run your first check

You ask hush to run a test or quality check on your project, and it wraps everything up neatly.

4
See the result
✅
All good

A simple checkmark appears, letting you move forward without extra clutter.

❌
Issue found

Only the most helpful error parts show up, so you can fix it quickly.

5
📝 Save your favorites

You create a short list of your go-to checks, so hush remembers your preferences.

6
🔄 Run a group of checks

You tell hush to handle several checks in a row, getting a quick overall summary.

🎉 Perfect workflow

Your project checks now run smoothly and efficiently, keeping everything tidy and saving lots of time for you and your AI helper.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is hush?

Hush is a Go-based command runner designed for coding agents, wrapping any shell command like pytest or npm test in a context-efficient layer. On success, it outputs just a single ✓ summary line with a label; on failure, it shows filtered error details while preserving the exit code. Use it via simple CLI like `hush "pytest -x"` or batch mode for multiple checks, with optional .hush.yaml configs for named runs like `hush lint` or `hush all`.

Why is it gaining traction?

It slashes token counts for AI agents by 96-99% on passing tests (e.g., pytest drops from 1k+ to 4 tokens), filtering verbose output from test runners, linters, and builds without losing debug info on fails. Batch mode with --continue handles pipelines like ruff + pytest, and flags like --grep "error" or --tail 30 make output precise. Unlike raw shell scripts, it's agent-optimized for hushed, actionable feedback.

Who should use this?

AI coding agent users scripting test/lint loops in tools like Cursor or Aider, where bloated command output kills context windows. Backend devs running Go/Cargo tests in agent-driven workflows, or Python teams batching ruff checks and pytest without token waste. Avoid if you need unfiltered logs for manual debugging.

Verdict

With 13 stars and 1.0% credibility score, hush is early-stage but polished—solid docs, benchmarks, and tests make it worth a spin for agent-heavy Go projects. Install via `go install` and benchmark your workflows; skip for mature prod pipelines.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.