Infrastructure scanning and dependency security. Built for engineers who run their own servers and ship their own code.
Run once, use forever. Scans your entire server and generates a persistent knowledge base. Claude auto-loads it every session — zero discovery overhead, instant context.
Multi-ecosystem supply chain attack scanner. Detects compromised packages, malware artifacts, and C2 connections across your entire dependency tree.
AI-scored GitHub repository discovery. Finds trending repos, scores them across 5 dimensions with LLM, and tweets the best ones daily.
8 design styles — from Swiss minimal to neo-brutalist trading terminals. Pick a style, Claude reads the DESIGN.md spec and transforms your dashboard. Your components stay, only the look changes.
One command scans every service, port, config, database, domain, SSL cert, and .env file. Single bash call, single permission prompt.
Creates a persistent CLAUDE.md + per-service docs. Auto-loaded every session. Claude knows your server without re-discovering anything.
Without this: 50-100K tokens burned on discovery per session. With this: zero. Claude reads the docs and starts working immediately.
Credentials in gitignored .credentials file. Optimization report flags security risks, performance issues, and expiring SSL certs.
80 lines max. Domain map, service table, architecture link. Everything Claude needs at a glance without wasting context.
One .md per service: nginx.md, postgresql.md, redis.md. TL;DR header, config tables, dependency graph. Read only when needed.
Critical / Important / Nice-to-have. Checks shared_buffers, SSL expiry, security headers, fail2ban, SSH config, disk usage.
# one-liner install (auto-updates every 6h)
curl -fsSL https://raw.githubusercontent.com/oopsalldev/server-scout/main/install.sh | bash
# run in claude code
/server-scout
# output
CLAUDE.md # lean index
.server-docs/services/*.md # per-service details
.server-docs/optimizations.md # recommendations
.server-docs/.credentials # passwords (gitignored)
JSON threat database with compromised versions, C2 IPs, malware hashes, and IOCs.
Checks node_modules, pip, Gemfile.lock, Cargo.lock, go.sum, composer.lock.
Malware files on disk. Active C2 connections. Malicious CDN refs. Bad lockfile shasums.
CLEAN / WARNING / COMPROMISED verdict. Remediation steps per threat.
# 1 — Clone the scanner (once per machine)
Downloads the repo so you can run the CLI locally.
git clone https://github.com/oopsalldev/npm-supply-chain-scanner
# 2 — Run a scan (from inside the clone)
Replace /your/project with the path you want checked (your app root).
cd npm-supply-chain-scanner
./scripts/scan.sh --path /your/project
# 3 — GitHub Actions (paste under your job’s steps: list)
Run after actions/checkout if the action needs your source tree. Prefer pinning @v… or a commit SHA instead of @main in production.
- uses: oopsalldev/npm-supply-chain-scanner@main
# Repo for stars & issues → github.com/oopsalldev/npm-supply-chain-scanner
# 4 — Claude Code (slash command in session)
Runs the scanner skill from your Claude Code workspace.
/scan
Scans GitHub every 30 minutes for trending repos across AI, security, devtools, and infrastructure. Sort by AI score, stars, latest discovery, or last push; filter by category and language.
LLM analyzes each repo across 5 dimensions: usefulness, code quality, documentation, innovation, and community. Weighted overall score 0-100.
Categorizes into ai-coding, ai-agents, llm-tools, security, web, devtools, infra, and more. Shareable URLs for every filter combination.
Top repos (score 70+) are automatically tweeted daily from @esenkalumut with AI-generated descriptions. No manual curation needed.
# browse — latest discoveries first
https://discover.oops.zone/?sort=latest
# category + language + sort (shareable)
https://discover.oops.zone/?category=ai-agents&lang=Python&sort=score
# last activity on GitHub (push date)
https://discover.oops.zone/?sort=updated
# follow for daily top repos
@esenkalumut on X8 design languages: Minimal (Swiss), Neo-Brutal (offset shadows), Glass (frosted blur), Soft (serif pastels), Cash App (fintech green), trAIder (lime-on-black terminal), Editorial (warm gray), Oops Default. Preview all live.
Each style has a DESIGN.md — exact colors, fonts, spacing, component patterns. Claude reads the spec and applies it to your existing components. No static HTML copying.
Your search still works. Your nav still works. Your buttons still work. Claude only changes how things look — colors, fonts, radius, shadows, spacing. Functionality untouched.
Every change tracked with git tags. One command to revert: git checkout oops-theme/pre-migration -- . Build, deploy, back to original.
Each style defined as a text specification — colors, fonts, patterns, anti-patterns. Claude reads the spec, not HTML. Your components stay, only visual properties change.
React, Next.js, Vue, Nuxt, SvelteKit, Astro, Angular, Remix, Solid.js, Vanilla. Tailwind v3 & v4. Monorepo support (Turborepo, Nx, pnpm).
Create a DESIGN.md in styles/{name}/ with colors, typography, component patterns. Add a preview HTML + screenshot. PR welcome — community-driven style library.
# one-liner install (auto-updates every 6h)
curl -fsSL https://raw.githubusercontent.com/oopsalldev/oops-theme/main/install.sh | bash
# run in claude code
/oops-theme
# preview styles
Built with Claude Code by oops.zone