← Back to Discover

Wiggum CLI

Wiggum CLI is an open-source command-line tool that scans a codebase, writes feature specs through an AI interview, and runs autonomous coding loops with Claude Code, Codex CLI or another CLI coding agent.

Coding agentsCommand-line tools
Visit website ↗

What it does

Wiggum is a command-line tool you install with npm and point at an existing project. Its init step maps the project's tech stack, structure and conventions. The new step then interviews you in plain English using that context and writes a spec as a markdown file that any CLI-based coding agent can read. The run step hands the spec to a coding agent such as Claude Code or Codex CLI and repeats plan, implement, test, verify and pull request phases with checkpoints. A terminal dashboard shows progress and token use while it works. Agent mode goes further. It works through a GitHub backlog by label and dependency order, creates a spec for each issue, reviews the diff against that spec and merges when checks pass. It can also run headless in CI pipelines. You bring your own API keys, and the site says paid Pro features are planned but not yet shipped.

Capabilities

  • Codebase scan that detects stack, structure and conventions
  • Spec generation through context-aware AI interviews
  • Autonomous plan, implement, test, verify and pull request loops
  • Works with Claude Code, Codex CLI or other CLI coding agents
  • Agent mode that processes a GitHub issue backlog
  • Headless mode for CI pipelines

When to use it

  • You want to turn a feature idea into a written spec before an agent writes code
  • You have a GitHub backlog you want an agent to work through with a review step
  • You already use a CLI coding agent and want it to run in repeatable loops