Skip to content

durang/claude-skills

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

41 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

[◠‿◠] Scan — AI Project Intelligence Engine

From idea to 100% launch. One command. Any project. Zero config.

What started as a progress tracker evolved into a complete project intelligence system for Claude Code — goal orchestration, verified task completion, security auditing, multi-language intent detection, and a living dashboard that never lets you lose context.

/track

That's it. It auto-detects your stack, runs your build, audits dependencies, measures git velocity, profiles your coding habits, tracks tasks with subtask ownership, hunts TODOs — and generates a branded, visual progress dashboard with honest launch readiness %.

╔═══════════════════════════════════════════════════════╗
║                                                       ║
║          Y  O  U  R  P  R  O  J  E  C  T             ║
║                                                       ║
║          [◠‿◠]  Scan · Progress Dashboard             ║
║                                                       ║
╚═══════════════════════════════════════════════════════╝

Install (10 seconds)

mkdir -p ~/.claude/skills/track && curl -sL https://raw.githubusercontent.com/durang/claude-skills/master/track/SKILL.md -o ~/.claude/skills/track/SKILL.md

Done. Open Claude Code in any project and type /track.


What You Get

Branded header with your project name

Every dashboard starts with your project name displayed large in spaced letters inside a clean box-drawing border. Your project feels like a product from day one.

Launch readiness with honest math

OVERALL        [████████████████░░░░]  80%  →  Production

Core Platform  [████████████████████] 100%  ✅ Ship-ready
Auth/Users     [████████████████████] 100%  ✅ Magic link + password
AI Chat        [████████████████████] 100%  ✅ DeepSeek V3 integrated
Payments       [████████████░░░░░░░░]  60%  ⚠️ Test mode only         ← BLOCKER
Security       [█████████████████░░░]  85%  🟡 2 items left
Deploy Prod    [██████████░░░░░░░░░░]  50%  🔴 Env vars + domain       ← BLOCKER

Blocker areas count 2x in the weighted average. Roadmap items are excluded. Your % reflects what actually matters for shipping — not optimism.

Active Tasks with ownership tracking

When you tell Claude "we need Gmail login" or "nos falta analytics", it creates a task with auto-generated subtasks tagged by who can do them:

### Gmail Login                              Auth · +5%
- [ ] Configure Google OAuth in Supabase       ← user
- [ ] Add provider env vars to Vercel          ← user
- [ ] Create Google sign-in button             ← auto
- [ ] Handle OAuth callback route              ← auto
- [ ] Update authState for Google provider     ← auto
- [ ] Test end-to-end                          ← user
> 0/6 · Next: needs OAuth config first
  • ← auto = Claude can execute it right now without your credentials
  • ← user = needs your dashboard access, API keys, or manual verification

When you say "let's go" / "sigamos" / "avancemos" — Claude reads the dashboard, finds all ← auto subtasks, and starts coding the highest-impact ones immediately. No menus. No "what would you like to do?" — just execution.

Visual charts from real git data

Daily commit activity:

                    Daily Commits (last 7 days)
  Tue 31  ████████████████████  20
  Wed 01  ██████               6
  Thu 02  ██████████████████   18
  Fri 03  █████                5
  Sat 04  ████                 4
  Sun 05  ███████              7
  Mon 06  █████                5   ← today

Your peak coding hours:

                    Peak Hours (when you code)
  4pm-8pm   ████████████████████  28 commits   ← peak
  midnight  ████████████████░░░░  21 commits
  9pm-11pm  ██████████████░░░░░░  16 commits

Codebase breakdown by language:

 TypeScript    ██████████████░░░░░░  14,498 lines
 TSX (React)   ████████████████████  22,669 lines
 CSS            █░░░░░░░░░░░░░░░░░░   1,648 lines
               ─────────────────────
 Total                                38,815 lines

Commit type distribution:

  fix     █████████████████████████████████████████████  45%
  feat    ████████████████████░░░░░░░░░░░░░░░░░░░░░░░░  20%
  chore   ████████████████████████████░░░░░░░░░░░░░░░░  35%

Progress timeline — your journey to launch:

                    Launch % Over Time
  Mar 26  ░░░░░░░░░░░░░░░░░░░░   0%   Project created
  Mar 29  ██████████░░░░░░░░░░  50%   Core + Auth
  Mar 30  ██████████████░░░░░░  70%   UI + AI Chat
  Apr 05  ███████████████░░░░░  75%   Security hardening
  Apr 06  ████████████████░░░░  80%   Vulns fixed + Tasks

Milestone map:

  ◉────◉────◉────◉────○────○────○────○
  10   25   50   75   85   90   95  100
                       ↑
                   you are here

Incremental scanning — saves tokens

The dashboard stores the last scanned commit hash. On re-scan:

Mode When What runs Savings
FULL First scan Everything Baseline
INCREMENTAL New commits since last scan Build, audit, velocity, health ~50%
SKIP Zero changes Nothing — just reports status ~95%

Architecture diagram

Auto-generated ASCII diagram showing your real project structure:

                    ┌──────────────────┐
                    │   Vercel (CDN)   │
                    └────────┬─────────┘
                             │
               ┌─────────────┴──────────────┐
               │    Next.js 16 App Router    │
               └──┬──────────┬──────────┬───┘
                  │          │          │
           ┌──────┘          │          └──────┐
    ┌──────┴──────┐   ┌──────┴──────┐   ┌──────┴──────┐
    │    Pages    │   │  API Routes │   │    Libs     │
    └──────┬──────┘   └──────┬──────┘   └──────┬──────┘
           │                 │                  │
    ┌──────┘    ┌────────────┼────────────┐     │
    ▼           ▼            ▼            ▼     ▼
 Zustand    Supabase    Lemon Squeezy  DeepSeek  Web Audio

Everything else

  • Build health — pass/fail, errors, warnings, route count
  • Dependency audit — outdated count, vulnerability scan
  • Code health — every TODO, FIXME, HACK with file:line
  • Security posture — headers, rate limiting, auth, CSRF, audit
  • Legal compliance — privacy, terms, disclaimers, GDPR
  • Business context — pricing, audience, model
  • Feature map — shipped, pending, roadmap
  • Logs — changes, decisions, learnings (append-only, never deleted)

Usage

# Auto-detect everything and scan
/track

# Tell it what you just did, then scan
/track Added Gmail authentication

# After a big session
/track Refactored API layer, added rate limiting, fixed 14 vulnerabilities

# Ask what's next — it reads the dashboard and tells you
que sigue?

Works With Everything

Stack Detection Checks
Node / Next.js / React / Vue package.json build, audit, outdated, routes
Rust Cargo.toml cargo check, cargo outdated
Go go.mod go build, modules
Python requirements.txt / pyproject.toml compileall, pip audit
Ruby Gemfile bundle audit
Java / Kotlin pom.xml / build.gradle build, deps
PHP composer.json composer audit
Monorepo Workspaces / Turborepo / Nx per-package + overall

Auto-detects: Vercel, Docker, AWS, Fly, Netlify, Railway, Supabase, Prisma, Firebase, GitHub Actions, GitLab CI, Jest, Vitest, Playwright, pytest.

How It Adapts

Project Type Areas Tracked
SaaS / Web App Core, Frontend, Auth, Payments, Security, Legal, SEO, Monitoring, Deploy
CLI / Library Core, API Design, Docs, Testing, CI/CD, Publishing
Mobile App Core, UI/UX, Auth, Push Notifications, App Store, Analytics
API / Backend Endpoints, Auth, Database, Rate Limiting, Docs, Deploy

FAQ

Does it change my code? No. It only reads your codebase and writes markdown status files.

Does it run my build? Yes — to verify build health. Uses your existing build command.

What if I have no dashboard yet? Creates one from scratch on first run. Zero setup.

Does it work in teams? Commit the dashboard to your repo. Everyone sees the same truth.

Can I use it on multiple projects? Yes. Installed globally at ~/.claude/skills/track/. Works in any directory.

Does it waste tokens on re-scans? No. Incremental mode only re-checks what changed since the last commit hash. If nothing changed, it skips entirely (~95% savings).

Alternative Install

# Clone for easy updates
git clone https://github.com/durang/claude-skills.git ~/claude-skills
ln -s ~/claude-skills/track ~/.claude/skills/track

# Update later
cd ~/claude-skills && git pull

Requirements

  • Claude Code (CLI, desktop, VS Code, JetBrains, or web)
  • A git repo
  • That's it

License

MIT


Built with Claude Code.

About

[◠‿◠] Scan — AI project intelligence engine for Claude Code. Zero-config progress dashboard, security audit, goal orchestration, verified task completion, multi-language intent detection. From idea to 100% launch.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages