ai-agent

Headlong

An open source agent microharness featuring persistent agency and recursive LLMs. Of bash, by bash, for bash; it's shells all the way down.
1,140 stars118 forksShellUpdated 9/9/2026100% free · open source
What it does

Headlong lets you write self‑modifying Bash agents that keep state across runs and can call LLMs recursively, all from the command line.

When to use it
  • You need a lightweight, always‑on automation that can remember prior interactions without a separate server.
  • You want to embed LLM‑driven decision‑making directly into existing Bash scripts or CI pipelines.
  • You prefer a pure‑shell solution that avoids Python/Node runtimes and can run on any Unix box.
Ready-to-paste prompt
./headlong run my‑agent "Create a weekly report of sales from the CSV files in data/ and email it to finance@example.com"
Heads up: Headlong requires Bash 5.0+ and a valid `OPENAI_API_KEY`; if the key is missing or the shell version is older, the agent will exit with an "API key not set" or "Unsupported Bash version" error.
Saves to your device
Use with Claude
New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install
Claude Code
mkdir -p ~/.claude/skills/headlong && curl -fsSL https://workflowstacks.com/api/skills/headlong/claude-skill -o ~/.claude/skills/headlong/SKILL.md
Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)
claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How Headlong works
Codeflow
Free to inspect

Headlong is a large Shell project (~29k lines across 154 code files, plus 111 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, Apache-2.0 license, has a test suite.

Size
Large codebase
~29k lines · 154 code files · ~4 h to skim
Setup
Developer setup
A real software project. Use it via its install path; don't expect to read it all.
Runs on
Shell
No API keys detected
Shell 58%Python 22%TypeScript 14%Rust 2%HCL 2%
Where to start reading
  1. 1
    README.md
    Start here — what it does and how to install it
  2. 2
    AGENTS.md
    The instructions the AI actually follows
  3. 3
    bin/llm
    Where the program starts running
  4. 4
    skills/README.md
    Inside skills/ — an example of what the AI is told to do
What's in each folder
skills/Prompts, skills & agent definitions10 files
web/Frontend / UI139 files
bin/Command-line entry points11 files
docs/Documentation4 files
deploy/Deployment / infrastructure75 files
telegram/Folder27 files
design/Proposals & design notes25 files
slack/Folder25 files
READMEHas testsDocumentedCI checksApache-2.0 licenseUpdated this month
Quick Actions
Details
Creator
laude-institute
Language
Shell
Category
ai-agent
Published
4/7/2026

Are you the creator of this tool? Claim your listing → and earn 85% of every sale.