OSCP+ Repos

Every repo on GitHub that mentions OSCP — hunted, cloned, read by AI, and judged by a human.

     

---

The Story

I was prepping for OSCP and wanted a single place to browse every relevant GitHub repo — notes, tools, writeups, cheatsheet.

So I built a pipeline that did the work for me.

Step 1 — Hunt. GitHub Search API, paginated across every repo created or pushed after mid-2025 with OSCP in its name or description. ~500 raw results.

Step 2 — Clone. Shallow-cloned (--depth 1) each repo locally. Repos too large (>500MB) fell back to the GitHub Contents API.

Step 3 — Extract. Walked every file tree, pulled READMEs. Stored everything as Base64 in SQLite — no flat files, no exports, just one clean .db.

Step 4 — Ask Claude. Fed each repo's description + file paths + README preview to claude-sonnet-4-6. Asked it to assign one of five categories:

Category	Meaning
OSCP+	Notes / cheatsheets exclusively for OSCP/PEN-200
Tools	Standalone security tools and scripts
Writeups	HTB / THM / CTF walkthroughs
Sensitive	OffSec exam content — Proving Grounds, exam reports
Generic	Multi-cert notes (OSCP + HTB + CRTP + ...)

Step 5 — Review. Manually went through every None result. Deleted 60+ repos that had nothing to do with offensive security — OS course projects in Chinese, Open Sound Control libraries, EV charging protocols, Oscar award prediction websites. You know how it is.

Result: 399 clean repos.

---

The Web App

Built a dark-mode dashboard on top of the SQLite DB — no data exports, no JSON files, the Node.js server queries the DB directly at request time.

Features

• Filter by category · search by name / description / language
• Sort by stars, size, last push, AI confidence
• Min-stars filter · per-page control
• Drag-to-resize columns
• Click any row → side panel with file tree, file list search, rendered README, AI reasoning
• Export current view as CSV

Stack: Node.js · Express · better-sqlite3 · Vanilla JS · Space Grotesk

---

Run Locally

git clone https://github.com/0xbugatti/oscp-collector
cd oscp-collector
npm install
node server.js
# → http://127.0.0.1:8787

The oscp.db file ships with the repo. No pipeline run needed to use the dashboard.

---

Pipeline (optional — to refresh the data)

pip install -r requirements.txt
export GITHUB_TOKEN=...
export ANTHROPIC_API_KEY=...
python main.py          # fetch → clone → extract → categorize → export
python main.py status   # check progress at any point

---

Deployed

oscpdb.vercel.app ← live app

---

Built while studying for OSCP. Try harder.