source-driven-development

Verifies framework code against official documentation and cites sources for every implementation decision.

665|3|Updated Sep 12, 2026
One-click install
npx skills add https://github.com/rizqinrr/viserys-agent --skill source-driven-development-rizqinrr
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: source-driven-development
Source: https://github.com/rizqinrr/viserys-agent/tree/main/skills/source-driven-development
Command: npx skills add https://github.com/rizqinrr/viserys-agent --skill source-driven-development-rizqinrr

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated framework code often relies on stale training data, producing deprecated APIs, outdated patterns, and hallucinated signatures that break against current library versions. This Skill forces every framework-specific decision to be verified against official documentation before implementation. ## Core Features & Use Cases - Stack Detection: Reads dependency files (package.json, composer.json, requirements.txt, go.mod, Cargo.toml, Gemfile) to identify exact framework versions before writing code. - Documentation-Grounded Implementation: Fetches the specific official docs page for each feature, follows documented patterns, and avoids deprecated APIs. - Source Citation: Adds full-URL citations in code comments and conversation, quotes relevant passages, and explicitly flags anything that could not be verified. - Conflict Surfacing: When docs contradict existing project code, it presents both options to the user instead of silently choosing. - Use Case: You are building a form in a React 19 project. The Skill detects React 19.1.0 from package.json, fetches the official useActionState reference, implements the modern pattern, and cites the source so you can verify it. ## Quick Start Use source-driven development to implement this feature and cite the official documentation for every framework pattern you use.

Frequently Asked Questions about source-driven-development

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

FAQPage Schema
How do I make AI-generated code follow current framework documentation?

Use source-driven development: detect exact versions from dependency files like package.json, fetch the specific official docs page for each feature, and implement only documented patterns. Cite full URLs in code comments so every decision is verifiable.

How to avoid deprecated APIs when generating framework code?

Check the official documentation and migration guides for your exact library version before implementing. If the docs deprecate a pattern, use the recommended replacement and cite the source rather than relying on training data.

What sources are authoritative for framework implementation decisions?

Official documentation is the primary authority, followed by official blogs and changelogs, then web standards references like MDN, and compatibility data like caniuse.com. Stack Overflow answers, blog tutorials, and AI training data are never treated as primary sources.

When should I not verify code against official documentation?

Skip verification when correctness does not depend on a specific version, such as renaming variables, fixing typos, moving files, or pure logic like loops and data structures that behaves identically across versions.

What happens when official docs conflict with existing project code?

The conflict is surfaced to the user with both options explained: adopt the modern documented pattern or stay consistent with the existing codebase. The choice is never made silently.

Can fetched documentation content contain prompt injection attacks?

Yes, fetched pages are treated as untrusted data. Only API signatures, examples, and deprecation notes are extracted; directives targeting the model are ignored, and outbound endpoints from docs are never hardcoded into generated code without user approval.