ABVX Agent Skills
ABVX Agent Skills is a portable execution layer for AI-native development: small, reviewable, validation-gated skillpacks for Codex-style project work.

What it is
ABVX Agent Skills turns repeatable AI-assisted work into reusable agent capabilities. It packages debugging, research, frontend and product builds, audits, handoffs, workbook forensics and workflow optimization into compact `SKILL.md` workflows with clear triggers, risk notes, attribution and validation gates.
The project sits inside the ABVX AI-native development ecosystem as a portable execution layer rather than a single interface. Each skillpack is designed to be small enough to review, explicit enough to invoke reliably and structured enough to support Codex-style project work without hiding the reasoning path.
The short version: portable, reviewable agent skillpacks for Codex-style project work: debugging, research, audits, frontend builds, handoffs and workflow orchestration.
The system is intended for:
- AI-assisted development workflows,
- project diagnostics and debugging,
- structured research and audits,
- frontend and product build passes,
- handoffs between agents or sessions,
- validation-gated execution in complex repositories.
ABVX Agent Skills belongs to the Workflow & Orchestration layer because it converts operational patterns into reusable procedures. It is connected to AGENTS.md Generator, SET, ABVX Lab and the broader ABVX development-system stack.






