npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2025 – Pkg Stats / Ryan Hefner

@boshu2/vibe-check

v1.8.1

Published

Track git commit patterns during AI-assisted coding (experimental)

Downloads

1,850

Readme


npx @boshu2/vibe-check

Focus: The Inner Developer Loop

vibe-check focuses on detecting the 4 Inner Loop Disasters — the immediate, developer-facing failures that happen during a coding session:

| Pattern | What Goes Wrong | How vibe-check Detects It | |---------|-----------------|---------------------------| | "Tests Passing" Lie | AI claims "fixed" but code doesn't work | Commits claiming success followed by immediate fixes | | Context Amnesia | AI forgets instructions from 5 minutes ago | Reverts, reimplementations, repeated similar fixes | | Instruction Drift | AI "improves" things you didn't ask for | Unrequested refactors, scope explosion | | Debug Loop Spiral | AI adds logging instead of fixing root cause | 3+ consecutive debugging commits |

These are the patterns you can catch and fix right now, in your current session.


The Problem

"AI can destroy months of work in minutes." — Gene Kim & Steve Yegge

AI-assisted coding creates a new failure mode: the debug spiral.

You ask AI for a fix. It breaks something else. You ask for another fix. It breaks more. Three commits later, you've lost 30 minutes and made negative progress.

vibe-check detects these spirals by analyzing your git history.


Quick Start

# Analyze recent work
npx @boshu2/vibe-check --since "1 week ago"

# Real-time spiral detection
npx @boshu2/vibe-check watch

# Visual dashboard
npx @boshu2/vibe-check dashboard

What It Measures

| Metric | Question | Target | |--------|----------|--------| | Trust Pass Rate | Does the code stick? | >95% | | Rework Ratio | Building or debugging? | <30% | | Debug Spirals | Are you stuck? | 0 active |


The 12 Failure Patterns

Vibe Coding identifies 12 failure modes across three developer loops. vibe-check focuses on the inner loop (where spirals form) while providing awareness of all 12.

Inner Loop (Minutes) — Primary Focus

| # | Pattern | Severity | vibe-check Detection | |---|---------|----------|----------------------| | 1 | Tests Passing Lie | 🔴 HIGH | Trust pass rate drops, fix commits follow "working" code | | 2 | Context Amnesia | 🟡 MEDIUM | Repeated fixes on same component, circular patterns | | 3 | Instruction Drift | 🟡 MEDIUM | Scope creep visible in commit message patterns | | 4 | Debug Loop Spiral | 🔴 HIGH | ✅ Core detection: 3+ fix commits on same component |

Middle Loop (Hours-Days) — Monitored

| # | Pattern | Severity | vibe-check Detection | |---|---------|----------|----------------------| | 5 | Eldritch Code Horror | 🔴 CRITICAL | High file churn, repeated touches to same files | | 6 | Workspace Collision | 🔴 HIGH | — (multi-agent, future) | | 7 | Memory Decay | 🟡 MEDIUM | Session baseline comparison shows degradation | | 8 | Multi-Agent Deadlock | 🔴 HIGH | — (multi-agent, future) |

Outer Loop (Weeks-Months) — Awareness

| # | Pattern | Severity | vibe-check Detection | |---|---------|----------|----------------------| | 9 | Bridge Torching | 🔴 CRITICAL | — (API tracking, future) | | 10 | Repository Deletion | 🔴 CRITICAL | — (git safety, future) | | 11 | Process Gridlock | 🔴 HIGH | — (CI/CD analysis, future) | | 12 | Stewnami | 🔴 HIGH | — (cross-repo, future) |

Inner loop mastery prevents middle/outer loop disasters. Catch the spiral at 3 commits, not 30.

See the full 12 Failure Patterns catalog for prevention strategies and AgentOps integration.


Watch Mode

Catches spirals as they form with personalized coaching:

VIBE-CHECK WATCH MODE
────────────────────────────────────────────────

  09:15  fix(auth): handle token refresh
  09:18  fix(auth): add retry logic
  09:22  fix(auth): increase timeout

  ⚠️  SPIRAL FORMING
      auth component, 3 fixes, 7 min

      Your history: 4 auth spirals, avg 18 min
      What worked: wrote a test (3x), took a break (1x)

      → Write a test before your next fix attempt.

────────────────────────────────────────────────

Watch mode learns from your history and suggests what's worked before.


Install

npm install -g @boshu2/vibe-check

Or use directly with npx:

npx @boshu2/vibe-check --since "2 hours ago"

Commands

Core Analysis

vibe-check                      # Analyze git history
vibe-check watch                # Real-time spiral monitoring
vibe-check dashboard            # Visual dashboard
vibe-check timeline             # View coding journey
vibe-check forensics            # Deep pattern analysis

Sessions

vibe-check start --level 3      # Quick session start
vibe-check session start        # Interactive session start
vibe-check session end          # End with metrics
vibe-check session status       # Check active session

Profile & Insights

vibe-check profile              # XP and achievements
vibe-check insights             # Your spiral patterns

Tools

vibe-check cache status         # Cache statistics
vibe-check cache clear          # Clear cache
vibe-check init-hook            # Install pre-push hook
vibe-check pipeline             # Audit CI/CD safety

Session Integration

Integrate with Claude Code or other AI dev tools:

# Start session
vibe-check session start --level 3

# End session - get JSON for logging
vibe-check session end --format json
{
  "metrics": { "trust_pass_rate": 92, "rework_ratio": 18 },
  "retro": {
    "failure_patterns_hit": [],
    "failure_patterns_avoided": ["Debug Spiral", "Context Amnesia"]
  },
  "inner_loop": {
    "health": "healthy",
    "issues_detected": 0,
    "tests_passing_lies": 0,
    "context_amnesia_incidents": 0,
    "instruction_drift_commits": 0,
    "debug_loop_detected": false
  }
}

When inner loop issues are detected:

⚠️ Inner Loop Health: WARNING

  🤥 2 "tests passing" lies detected
  🎯 Instruction drift: 3 commits went outside scope

  → STOP: AI claimed success but code needed fixes. Verify builds/tests pass.

For Autonomous Agents

vibe-check measures human-AI collaboration sessions—the inner loop where you're working with an AI assistant.

For autonomous agents that run without a human in the loop, see 12-Factor AgentOps—DevOps and SRE patterns adapted for AI systems.

| Use Case | Tool | |----------|------| | Working with AI (pair programming) | vibe-check | | Running AI autonomously (agents) | 12-Factor AgentOps |


Philosophy

vibe-check is a mirror, not a judge.

It answers one question: are you building, or are you spiraling?

Use it for self-reflection. Catch your own patterns. Improve your own flow.

It's not a productivity metric. It's not for performance reviews.


Requirements

  • Node.js >= 20.0.0
  • Git repository with commits

Learn More


License

MIT