snarktank/ai-dev-tasks
A minimal three-file task loop for AI dev agents: create a PRD, generate a task list, then execute one task at a time with human check-ins. Agentiquette scores it 63/100 (B-tier), methodology v1.0, scored 2026-07-07.
Key facts
| Category | Workflow / loop tool |
|---|---|
| License | Apache-2.0 |
| Maintainer | snarktank (Ryan Carson) |
| Compatible agents | cursor, claude-code, generic |
| File patterns | |
| Triggering | manual-reference |
| GitHub metrics | 7,777 stars · 1,729 forks · 10 open issues (fetched 2026-07-07) |
| Last commit | 2025-11-05 |
| Created | 2025-04-19 |
Editorial analysis
The minimalist counterpoint to every framework in this index: three markdown files that encode plan-then-execute-one-task-at-a-time, and nothing else. As pedagogy it is excellent; the one-task discipline it teaches is the part of loop reliability most teams skip. Verification remains the human's job, by design.
Risks
Eight months quiet at fetch; as three markdown files, it ages more gracefully than code, but conventions drift.
Evidence notes: Leverage: near-zero adoption cost for a real discipline change. Reliability: human-verified by design; nothing automated. Maintenance: 8-month gap; markdown decays slowly.
Install and first run
Copy the three markdown files into your project and follow the workflow.Alternatives in this category
An AI task-management loop that decomposes a spec into ordered, dependency-aware tasks and drives an agent through them one at a time across…
Spec-driven development workflow for Claude Code: requirements, design, tasks, and implementation as gated phases with approval between…
Compare with claude-task-master · Compare with claude-code-spec-workflow
Embed this badge
Maintainers of scored entries can embed their live badge; it updates when the score does and links back to this page.
[](https://agentiquette.com/index/repos/ai-dev-tasks)Frequently asked questions
Is snarktank/ai-dev-tasks good?
snarktank/ai-dev-tasks scores 63/100 (B-tier) on Agentiquette's methodology v1.0, scored 2026-07-07. Strengths and weaknesses are in the editorial analysis; the score is explained dimension by dimension in the scorecard.
What is snarktank/ai-dev-tasks for?
A minimal three-file task loop for AI dev agents: create a PRD, generate a task list, then execute one task at a time with human check-ins.
Is snarktank/ai-dev-tasks maintained?
Last commit 2025-11-05, 10 open issues at fetch time (2026-07-07). Maintenance is scored at 35 of 100.
Maintainer of snarktank/ai-dev-tasks? Claim this listing and suggest corrections.