majiayu000Community·9 Agent Skills Included

vibeguard

Real-time guardrails that block hallucinated and unsafe code changes

Blocks AI coding agents from shipping hallucinated APIs, duplicate files, and unverified changes before they land. Intercepts dangerous shell commands, force pushes, and destructive cleanup in real time through hooks and static guards. Returns a concrete fix instruction with every interception so agents self-correct instead of failing silently. Works with Claude Code, Codex CLI, and Gemini CLI from the first prompt after install.

37|6

↗

npx skills add majiayu000/vibeguard --all -g -y

Available:

Gives the AI agent the repository-wide operating rules for VibeGuard, including scope limits, safety policies, validation commands, and which directory guides to read before changing hooks, rules, or runtime code.

All Skills in This Repository (9)

Pure Emerald Level Indicators

📦 In Repo

majiayu000

plan-folw

Analyze codebases for redundant design patterns and generate step-by-step refactoring plans.

auto-optimize

Automate project analysis, evaluation, design, and optimization with VibeGuard guards.

fixflow

Orchestrate software development with planning, sequential implementation, testing, and commit strategies.

optflow

Automate discovery, prioritization, and step-by-step execution of code optimization tasks.

plan-mode

Generate phased technical execution plans with risks and YAML frontmatter.

iterative-retrieval

Iteratively retrieve codebase information using Glob, Grep, and AST searches.

eval-harness

Quantify AI-generated code quality using pass@k and pass^k metrics.

vibeguard

Prevent LLM code generation hallucinations with a seven-layer defense framework.

strategic-compact

Compress long AI conversation context by retaining critical decisions and incomplete tasks.

Community

Basic

Frequently Asked Questions

FAQPage Schema

How to install VibeGuard?▼

Run `npx skills add majiayu000/vibeguard --all -g -y` in your terminal to install all skills in this suite globally.

What does VibeGuard protect against?▼

It blocks hallucinated APIs, duplicate files, hardcoded secrets, dangerous commands like rm -rf and force pushes, and unverified 'done' claims from AI coding agents.

Does VibeGuard work with Claude Code and Codex?▼

Yes. It installs native rules and real-time hooks for Claude Code, Codex CLI, and Gemini CLI, and protection is active from the first prompt of a new session.

How does VibeGuard stop bad AI code changes?▼

It combines rule injection at session start, pre- and post-edit hooks that block or warn in real time, and static guards that scan for duplicates and code-quality issues, each returning a fix instruction.

Can I monitor VibeGuard hook activity?▼

Yes. It ships local observability commands for dashboards, hook health, stats, and metrics export, all running locally without external telemetry.