commit 414cad40c5669e58ece210b29aac97c05443a9d0 Author: Stephen Groundwater Date: Wed Oct 14 21:18:44 2026 -0400 Initial structure for AI and security practice notes Co-Authored-By: Claude Opus 5.5 diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..73d36c0 --- /dev/null +++ b/.gitignore @@ -0,0 +1,6 @@ +.env +*.kdbx +*.key +*.pem +credentials* +.obsidian/workspace*.json diff --git a/CLAUDE.md b/CLAUDE.md new file mode 100644 index 0000000..4bc5a6d --- /dev/null +++ b/CLAUDE.md @@ -0,0 +1,17 @@ +# CLAUDE.md: ai-security-practice + +Stephen's personal knowledge base for AI and security automation. Read README.md for the line between his knowledge (belongs here) and SANS courseware (never here). + +## Rules for AI sessions in this repo + +- **Never add SANS courseware content**: no book/workbook/slide text, close paraphrases, lab walkthroughs, lab answers, or verbatim workbook commands. If Stephen pastes something that looks like courseware, say so and ask him to restate it in his own words. +- Write entries from what Stephen tells you in conversation, in his framing. Ask follow-up questions that deepen his understanding instead of filling gaps from the course. +- Connect concepts to his real infrastructure: Gluebox Drupal/LAMP servers on Linode, CrowdSec (gluebox-crowdsec repo), Gitea at code.gluebox.com, rsvp-system, the Whitingham client. +- Public vendor docs (Tines, Open WebUI, AWS, Azure, AbuseIPDB, VirusTotal, LiteLLM) are fair game; cite URLs. +- Never commit secrets (API keys, tenant IDs, lab credentials). Those belong in KeePassXC. + +## Conventions + +- Journal: `journal/YYYY-MM-DD.md`, using the sections in `journal/_template.md` +- Concepts: `concepts/kebab-case-name.md`, linked from journal entries +- Remote: `gitea-gluebox:tux/ai-security-practice.git` (SSH; HTTPS is blocked by Cloudflare) diff --git a/README.md b/README.md new file mode 100644 index 0000000..cdb4cf2 --- /dev/null +++ b/README.md @@ -0,0 +1,28 @@ +# ai-security-practice + +My own knowledge base for AI and security automation: concepts in my words, applied to +Gluebox infrastructure, plus the workflows and code I build. It started with SANS SEC598 +(UPenn partnership, October 2026) and is meant to keep growing after that. + +## The line + +This repo holds **what I understand and what I build**, never course material. + +| ✅ Belongs here | ❌ Never here | +|---|---| +| Concepts written from memory, in my words | Book, workbook or slide text, even paraphrased | +| How an idea applies to Gluebox, my servers or clients | Lab step-by-steps or lab answers | +| Code, Tines stories and configs I wrote | Screenshots of courseware | +| Errors I hit and how I fixed them | Page-by-page summaries | +| Questions to explore; links to public vendor docs | Commands copied verbatim from the workbook | + +Test: *could I write this note without the book open?* If not, it goes in the private, +non-AI notes (`~/sans-courseware/notes/`), not here. + +## Layout + +- `journal/`: one file per day (`YYYY-MM-DD.md`): what clicked, what's still confusing, ideas +- `concepts/`: one file per concept, refined over time +- `applied/`: plans for using these ideas on Gluebox, rsvp-system and client work +- `builds/`: things I've built (scripts, Tines story exports, prompts, configs) +- `questions.md`: open questions to chase down diff --git a/builds/.gitkeep b/builds/.gitkeep new file mode 100644 index 0000000..e69de29 diff --git a/journal/_template.md b/journal/_template.md new file mode 100644 index 0000000..56e02a0 --- /dev/null +++ b/journal/_template.md @@ -0,0 +1,13 @@ +# YYYY-MM-DD: Day N, + +## What clicked + +## Still fuzzy + +## Where this applies to my work + +## Built / tried + +## Problems and fixes + +## Follow-ups diff --git a/questions.md b/questions.md new file mode 100644 index 0000000..f60b255 --- /dev/null +++ b/questions.md @@ -0,0 +1,3 @@ +# Open questions + +- [ ] Ask the instructor (day 1): is it OK to use AI with my own notes, written in my own words?