feature-pipeline

Orchestrates multi-phase feature development with checkpoint gates and parallel agent streams.

4|2|Updated Jan 27, 2026
One-click install
npx skills add https://github.com/Arete-Consortium/ai-skills --skill feature-pipeline-arete-consortium
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: feature-pipeline
Source: https://github.com/Arete-Consortium/ai-skills/tree/main/workflows/feature-pipeline
Command: npx skills add https://github.com/Arete-Consortium/ai-skills --skill feature-pipeline-arete-consortium

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Feature development without structure leads to implementation before requirements are understood, tests written after code, and security reviewed never. This Skill enforces a discovery-then-implement-then-deliver sequence with mandatory human checkpoints, parallel agent execution, and numbered artifacts so no phase operates without the output of its predecessor. ## Core Features & Use Cases - Phased Pipeline with Checkpoint Gates: Runs Discovery, Implementation, Review, and Delivery phases, each requiring explicit user approval before proceeding. - Parallel Agent Streams: Spawns backend, frontend, and test agents in parallel during implementation, plus security and performance review agents before delivery. - Numbered Artifact Chaining: Produces ten ordered artifacts (01-requirements.md through 10-final-checklist.md) in a .feature-dev/ directory, with a state.json tracking phase status and git refs for clean per-phase revert. - Use Case: When building a multi-layer feature like user authentication, use this workflow to gather requirements, design architecture, implement backend and frontend in parallel, run security and performance reviews, and generate deployment config and documentation with full traceability. ## Quick Start Ask Claude to run the feature-pipeline workflow to build your feature, starting with the discovery phase to gather requirements and architecture before any code is written.

Frequently Asked Questions about feature-pipeline

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a multi-phase feature development workflow with AI agents?▼

Start the feature-pipeline workflow, which runs Discovery, Implementation, Review, and Delivery phases in sequence. Each phase produces numbered artifacts and requires your explicit approval at checkpoint gates before the next phase begins.

How does the workflow coordinate parallel backend and frontend agents?▼

During Phase 2, backend and frontend agents run in parallel using the requirements and architecture artifacts as input, while the test agent starts after implementation completes. A maximum of three parallel agents per phase prevents context fragmentation.

When should I not use a feature development pipeline workflow?▼

Avoid it for backend-only or frontend-only changes, simple CRUD endpoints, exploratory prototyping, or projects lacking established conventions. The discovery phase and checkpoints add overhead that only pays off for multi-layer features.

What happens when a security review finds critical issues?▼

Critical and high severity findings block the delivery phase entirely. They must be remediated and re-reviewed before Checkpoint 2 can be approved, while medium and low findings are logged as follow-up items.

Can I revert a single phase of the feature pipeline?▼

Yes, each phase records git_ref_start and git_ref_end in state.json. Revert a phase with git revert --no-commit over that range, producing a single revert commit without rewriting history.