Gald3r

Quickstart

Install gald3r in any project and run your first commands in under 10 minutes.

View on GitHub Wiki

Prerequisites

  • • Python 3.11+ with uv installed (pip install uv)
  • • Any supported AI coding tool: Cursor, Claude Code, Gemini CLI, Codex, OpenCode, or Copilot
  • • A git-initialized project folder
1

Install gald3r

uv run gald3r install

Run from any project root. gald3r creates a .gald3r/ folder with TASKS.md, PROJECT.md, PLAN.md, CONSTRAINTS.md, BUGS.md, and an initial config. No database, no cloud, no backend required.

2

Open your IDE of choice

gald3r works with Cursor, Claude Code, Gemini CLI, OpenAI Codex, GitHub Copilot, and 30+ other platforms. The same .gald3r/ folder is read by all of them — switch tools mid-task without losing context.

3

Create your first task

@g-task-add "Your feature or bug here"

This creates a task spec file in .gald3r/tasks/ with acceptance criteria, status history, and metadata. Tasks are plain markdown — readable by agents and humans alike.

4

Implement against the spec

@g-go-code

The g-go-code pipeline reads the task spec, implements against the acceptance criteria, and gates completion behind a fresh independent reviewer. The implementing agent cannot mark its own work done.

5

Check what's open

@g-status

Shows active tasks by priority, blocked items, recent completions, and next recommended actions. Use @g-report for a full markdown report you can share with your team.

What gets installed

.gald3r/
├── TASKS.md          ← active task index
├── PROJECT.md        ← mission, goals, project linking
├── PLAN.md           ← master strategy and roadmap
├── CONSTRAINTS.md    ← guardrails agents must never violate
├── BUGS.md           ← bug registry (zero-ignore policy)
├── config/           ← agent config, sprint, KPI settings
├── tasks/            ← individual task spec files
└── bugs/             ← individual bug spec files

.cursor/  (or .claude/, .gemini/, etc.)
└── rules/, skills/, agents/, commands/, hooks/

All files are plain markdown. No database. No cloud. Commit them to git — your task history is version-controlled alongside your code.

Explore Commands →How gald3r works →