Pu.sh: 400-Line Shell Script Coding-Agent Harness from HN

Pu.sh is a full coding-agent harness written in ~400 lines of shell script, designed for maximum portability. It requires only sh, curl, and awk — no npm, pip, Docker, or other dependencies. The project is inspired by Pi (pi.dev) and uses the same 7-tool surface and exact-text edit model.
Key Features
- Supported models: Anthropic (Claude) and OpenAI
- 7 built-in tools: bash, read, write, edit, grep, find, ls
- REPL mode for interactive use
- Auto-compaction of conversation context
- Checkpoint/resume capabilities
- Pipe mode for scripting
- 90 no-API tests to verify functionality without external calls
Notably missing: TUI, streaming, image support, OAuth, Windows compatibility — and, as the author puts it, "dignity."
Code and Philosophy
The author admits they cannot read most of the awk code, which was written by Pi/Claude/Codex. The reasoning loop handles JSON parsing and tool calls across turns using awk. The agent loop itself is tiny — most of the complexity in real agent CLIs comes from DX and hardening. The project is released under a self-imposed rule: no new dependencies and under 500 LOC.
Setup
curl -sL pu.dev/pu.sh -o pu.sh && chmod +x pu.sh
./pu.sh
All you need is curl, awk, and an API key. The author encourages readers to build their own harness adapted to their needs.
📖 Read the full source: HN AI Agents
👀 See Also

Upfront: A Claude Code Plugin That Forces Thinking Before Coding
Upfront is a Claude Code plugin with 20 skills that challenges developers before generating code. It uses three commands: /upfront:feature to push back on vague requirements, /upfront:plan to break work into ~400 LOC phases, and /upfront:build to execute with TDD and review per phase.

x402 API Gateway for OpenClaw Bots: One Endpoint Replaces 18 API Keys
An x402 API gateway eliminates the need for multiple API keys in OpenClaw bots by providing access to 18 services including smart LLM routing, web search, maps, travel, food, AI, and finance data through a single endpoint authenticated via USDC wallet credits.

Claude Code Session Data Loss: Backup Script for Windows & Mac
Users report silent session data loss in Claude Code. Here's a free, automated backup script for Windows and Mac using PowerShell and launchd.

Qwen3.6-27B as a Local Reasoning Layer: 2-Week Multi-Agent Test Results
A developer replaced Claude with local Qwen3.6-27B in a multi-agent orchestrator for 2 weeks across 47 workflows. Results: strong reasoning, 12% tool-call format errors, and 12k token context limit.