Cyberlium

AI › Module 9 › Lesson 4

BeginnerModule 9Lesson 4/5

Lab — Findings

Findings lab on `$AI_LAB` — writeup, severity backlog, responsible disclosure checklist bundled.

25 min+40 XP3 quiz
Module progress4 of 5

Visual · t32_findings_lab

Lab: findings pack from YOUR $AI_LAB eval artifacts. Original Cyberlium.

Opening

Ship professional AI findings artifacts from YOUR authorized eval — writeup, triage, ethics — zero jailbreak kits or stranger app evidence.

From `$AI_LAB` labs Modules 6–8: (1) report cover — scope, RoE, analyst, UTC; (2) one complete finding writeup Module 9-1; (3) prioritized backlog five findings Module 9-2; (4) responsible disclosure checklist Module 9-3; (5) remediation summary stub linking app/eval/governance rows; (6) limitations paragraph (lab vs prod, brief gaps); (7) integrity — `$AI_LAB` only, no stranger app test, no jailbreak kit repro, no prod paste; (8) chmod 600 index.md. Professional tone — mentor-readable. Cross-link LLM app, secure dev, and governance packs.

1. Lab deliverables

Writeup + priority backlog + responsible disclosure — single findings thread.

Limitations paragraph mandatory — honest lab scope.

Command guide

Try these commands — Lab deliverables

═══ 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/ MITRE ATLAS — https://atlas.mitre.org/ NIST AI RMF — https://www.nist.gov/itl/ai-risk-management-framework

═══ 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/findings/findings-lab-index.md" <<'EOF'
# Findings Lab Index — YOUR authorized tests
- finding-writeup-template.md (copy per finding)
- severity-triage.md
- responsible-ai-disclosure.txt
- Link injection lab: bot/echo_bot.py test results
- Map each finding to OWASP LLM01–LLM10
## Integrity
- All findings from YOUR lab toys or YOUR staging apps only
EOF

Command — copy this

grep -E 'OWASP LLM|Integrity|YOUR' "$AI_LAB/findings/findings-lab-index.md"
ls "$AI_LAB/findings/"

═══ WINDOWS ═══

Command — copy this

Get-Content $HOME/cyberlium-lab/t32-ai/findings/responsible-ai-disclosure.txt | Select-String NEVER

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/); MITRE ATLAS (https://atlas.mitre.org/); NIST AI RMF (https://www.nist.gov/itl/ai-risk-management-framework). Run every command in the box — install first, then the usage lines — only on YOUR lab / program scope.

2. Quality bar

Each finding traceable to Module 6–8 evidence row.

Zero jailbreak kits or stranger app refs in any section.

3. Integrity scan

Scan for stranger refs, bypass kits, prod paste — zero hits.

Practice label if brief requires.

4. What you ship: findings lab pack

Writeup + triage + ethics + index + integrity — chmod 600.

5. What you record before the next lesson

Findings lab pack 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

    Findings lab includes public jailbreak kit appendix and stranger app scan export.

  • Right

    Findings lab pack on `$AI_LAB`. Next: quiz.

Mission: findings lab

1) Finding writeup complete. 2) Priority backlog five rows. 3) Responsible disclosure checklist. 4) Integrity scan; chmod 600.

Stuck? Ask Cyberlium AI Mentor

Ask Mentor: “Minimum writeups if brief sparse?”

Knowledge Check

1

APPLY: Findings lab includes:

Multiple choice

Knowledge Check

2

APPLY: True or False: Hide limitations in findings lab.

True or False

Knowledge Check

3

APPLY: Finding writeup repro:

Multiple choice

← Previous

Answer all 3 knowledge checks to continue. (0/3 answered)