Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

PRD TaskMaster (Atlas Engine)

A zero-config engine that turns any goal into a validated PRD and an executable TaskMaster task graph.

AutomationAdvanced★ 586⑂ 58AI score 7/10Last updated: Jul 9, 2026

What it does

prd-taskmaster (the Atlas engine) is a four-phase pipeline that takes you from a one-line goal to a ready-to-execute plan:

  • Phase 0 — Backend resolution: prefers the atlas-engine MCP server, falls back to the bundled zero-dependency script.py.
  • Phase 1 — Preflight: auto-detects TaskMaster, AI providers (Claude Code / Codex CLI / API keys) and research routes. Asks zero setup questions.
  • Phase 2 — Discovery: adaptive, one-question-at-a-time brainstorming that works for any domain.
  • Phase 3 — Generate & Validate: writes the spec, runs quality validation (fails on TBD/TODO/placeholder text), parses it into tasks.json and enriches task metadata.
  • Phase 4 — Handoff: recommends exactly one execution mode — Verified Loop, Auto-Execute, or Plan & Drive.

No TaskMaster installed? "Native Mode" (--manual) authors a TaskMaster-compatible tasks.json directly and validates it with validate-tasks + enrich-tasks.

Who it's for

  • Solo devs and small teams who stall on writing PRDs and breaking work down
  • Anyone already running a task-driven development loop with TaskMaster
  • Planners who need domain-agnostic specs (software, pentest, business, learning)
  • Power users who want parallel research fan-out across multiple subagents

Examples

  1. New SaaS feature: "I want to build a dashboard with subscription billing" → discovery Q&A → .taskmaster/docs/prd.md → validation passes → ~20 tasks with subtasks → handoff to Verified Loop.
  2. Machine without TaskMaster: run /atlas --manual; the skill writes tasks.json itself, then verifies it via validate-tasks and enrich-tasks --require-phase-config.
  3. Parallel research: parallel-plan --missing-only emits research packets, N subagents work them in parallel, and parallel-apply merges everything in a single atomic write plus a complexity report.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/anombyte93/prd-taskmaster/HEAD/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the SKILL.m folder from the GitHub repo anombyte93/prd-taskmaster into my ~/.claude/skills/prd-taskmaster/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/anombyte93/prd-taskmaster.git ~/.claude/skills/prd-taskmaster

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal and make sure the skills folder exists: mkdir -p ~/.claude/skills
  2. Clone the repository into place: git clone https://github.com/anombyte93/prd-taskmaster.git ~/.claude/skills/prd-taskmaster
  3. Verify the key files landed: ls ~/.claude/skills/prd-taskmaster (you should see SKILL.md, script.py, and a phases/ folder).
  4. Confirm Python 3 is available: python3 --version — required for the CLI fallback backend.
  5. (Optional) Install TaskMaster for the full experience: npm i -g task-master-ai
  6. Restart Claude Code, then invoke the skill with /atlas, /prd-taskmaster, or simply "I want to build …".
  7. Check the single Engine backend: ... line printed on the first run to see whether you're in MCP or script.py mode, then answer the discovery questions.
View source on GitHub ↗License: MIT