AI › Module 10 › Lesson 1
Checklist
Topic 32 capstone checklist — AI mindset, injection, poisoning, threats, LLM apps, secure dev, governance, findings on `$AI_LAB`.
Visual · t32_capstone_checklist
Checklist consolidates AI Security literacy pillars. $AI_LAB only. Original Cyberlium.
Opening
Capstone proves AI Security literacy on YOUR authorized `$AI_LAB` work — not jailbreak kits or stranger app testing.
Assemble checklist covering Modules 1–9: AI mindset and lab toys, LLM basics, prompt injection defenses, poisoning/adversarial literacy, AI-powered threat literacy, LLM app security Top 10/RAG/output, secure SDLC eval/secrets, governance/red team/RoE, findings and responsible disclosure. Use on capstone walk of YOUR `$AI_LAB` — never jailbreak weaponization, never stranger app testing, never prod LLM without RoE, never fraud deepfake kits. Each row links artifact or honest N/A. Notes chmod 600 under $HOME/cyberlium-lab/t32-ai/. Integrity rows: `$AI_LAB` only, YOUR apps authorized red team, no weaponized kits, limitations documented.
1. Scope and ethics
$AI_LAB named with RoE reference. NEVER jailbreak weaponization, NEVER stranger app testing, NEVER fraud deepfake kits, NEVER prod without ticket.
AI red team = authorized eval of YOUR apps — not public bypass distribution.
Command guide
Try these commands — Scope and ethics
═══ TOOLS & WEBSITES ═══ Browse / read these (authorized learning only — stay in YOUR lab / program scope)
OWASP LLM Top 10 — https://owasp.org/www-project-top-10-for-large-language-model-applications/ NIST AI RMF — https://www.nist.gov/itl/ai-risk-management-framework MITRE ATLAS — https://atlas.mitre.org/ CISA AI — https://www.cisa.gov/ai
═══ INSTALL ═══
Linux (Debian/Ubuntu):
Command — copy this
sudo apt install python3
macOS:
Command — copy this
brew install python3
Windows: Download https://python.org/downloads/
═══ LINUX / macOS ═══
Command — copy this
export AI_LAB=${AI_LAB:-$HOME/cyberlium-lab/t32-ai}
cat > "$AI_LAB/notes/capstone-checklist.txt" <<'EOF'
[ ] lab.env + notes/ bot/ injection/ threats/ owasp/ eval/ governance/ findings/
[ ] roe.txt + no-weaponized-ai.txt reviewed
[ ] LLM basics + trust boundaries + context/tools literacy
[ ] echo_bot.py + injection tests on YOUR bot only
[ ] poisoning/adversarial named risks + supply chain checklist
[ ] phishing/deepfake/malware detection signal notes (no payloads)
[ ] OWASP LLM Top 10 index + RAG/output checklists
[ ] secure SDLC + eval checklist + secrets-in-prompts
[ ] governance + red team named + policy-and-roe
[ ] finding templates mapped to OWASP LLM categories
[ ] never: malware gen / deepfake fraud / jailbreak stranger SaaS / stranger API scrape
EOFCommand — copy this
grep '\[ \]' "$AI_LAB/notes/capstone-checklist.txt"
python3 -c "print('Capstone: AI Security — YOUR t32-ai lab only')"Primary tools to practice this lesson: grep, python3. Reference sites: OWASP LLM Top 10 (https://owasp.org/www-project-top-10-for-large-language-model-applications/); NIST AI RMF (https://www.nist.gov/itl/ai-risk-management-framework); MITRE ATLAS (https://atlas.mitre.org/); CISA AI (https://www.cisa.gov/ai). Run every command in the box — install first, then the usage lines — only on YOUR lab / program scope.
2. Literacy pillars
Injection, poisoning, threats, LLM apps, secure dev, governance, findings — linked rows.
Cross-link Topic 31 Cloud Security as separate track — unauthorized paths refused.
3. Teardown
Notes secured chmod 600, report pack ready, lab keys rotated stub per brief.
Archive retention note for mentor handoff.
4. What you ship: Topic 32 capstone checklist
One-page AI Security literacy capstone checklist for $AI_LAB.
5. What you record before the next lesson
Checklist file path.
6. Wrong vs right: stranger SaaS vs YOUR toy LLM
Worked failure — same MSF word, opposite target. Right never needs a café Wi-Fi or classmate laptop.
Wrong
Checklist includes row 'jailbreak weaponization kit for capstone bonus.'
Right
Checklist written. Next: Lab Path.
Mission: write capstone checklist
1) Ethics/scope section. 2) Modules 1–9 pillar rows. 3) Teardown and handoff section.
Stuck? Ask Cyberlium AI Mentor
Ask Mentor: “Top five capstone artifacts?”
Knowledge Check
APPLY: Checklist includes:
Multiple choice
Knowledge Check
APPLY: True or False: Jailbreak weaponization required for capstone.
True or False
Knowledge Check
APPLY: Teardown includes:
Multiple choice