Skip to content

ShulkwiSEC/bb-huge / skills/curated/ai-jailbreak-system-prompts

ai-jailbreak-system-prompts

Unverified

Advanced techniques for bypassing LLM safety filters, instruction tuning, and system prompt restrictions using specialized linguistic constructs, hypothetical scenarios, and persona adoption.

FileKindSize
SKILL.mdskill md3.6 KB
evals/evals.jsonconfig540 B
scripts/process.pyscript7.8 KB
---
name: ai-jailbreak-system-prompts
description: >
  Advanced techniques for bypassing LLM safety filters, instruction tuning, and system prompt 
  restrictions using specialized linguistic constructs, hypothetical scenarios, and persona adoption.
domain: cybersecurity
subdomain: ai-red-teaming
category: Prompt Engineering
difficulty: advanced
estimated_time: "2 hours"
mitre_atlas:
  tactics: [AML.TA0001]
  techniques: [AML.T0043, AML.T0051]
mitre_attack:
  tactics: [TA0009]
  techniques: [T1592]
platforms: [ai, web]
tags: [ai, jailbreak, prompt-engineering, llm, safety-bypasses, red-teaming]
tools: [chat-interfaces, burp-suite, custom-scripts]
version: "1.0"
author: CyberSkills-Elite
license: Apache-2.0
---

# AI Jailbreaking & System Prompt Bypasses

## When to Use
- When conducting security assessments of Large Language Models (LLMs) integrated into chatbots, virtual assistants, or backend AI data processing pipelines.
- To demonstrate how instruction-tuned models can be forced into producing harmful, unethical, or restricted outputs by carefully crafting adversarial prompts.


## Prerequisites
- Access to target AI/ML system or local model deployment for testing
- Python 3.9+ with relevant ML libraries (transformers, torch, openai)
- Understanding of LLM architecture and prompt processing pipelines
- Authorized scope and rules of engagement for AI red team testing

## Workflow

### Phase 1: Understanding Target Model Constraints

```text
# Concept: LLM safety filters ```

### Phase 2: Persona Adoption Attacks

```text
# ```

### Phase 3: Developer Mode & Fictional Scenarios

```text
# ```

### Phase 4: Payload Encoding & Obfuscation

```text
# ```

#### Decision Point πŸ”€
```mermaid
flowchart TD
    A[Craft Prompt ] --> B{Bypass Successful ]}
    B -->|Yes| C[Capture Output ]
    B -->|No| D[Refine Approach ]
    C --> E[Test Edge Cases ]
```

## πŸ”΅ Blue Team Detection & Defense
- **Filter Ensembling**: **Context Monitoring**: Key Concepts
| Concept | Description |
|---------|-------------|
## Output Format
```
Ai Jailbreak System Prompts β€” Assessment Report
============================================================
Target: [Target identifier]
Assessor: [Operator name]
Date: [Assessment date]
Scope: [Authorized scope]
MITRE ATT&CK: [Relevant technique IDs]

Findings Summary:
  [Finding 1]: [Severity] β€” [Brief description]
  [Finding 2]: [Severity] β€” [Brief description]

Detailed Results:
  Phase 1: [Phase name]
    - Result: [Outcome]
    - Evidence: [Screenshot/log reference]
    - Impact: [Business impact assessment]

  Phase 2: [Phase name]
    - Result: [Outcome]
    - Evidence: [Screenshot/log reference]
    - Impact: [Business impact assessment]

Risk Rating: [Critical/High/Medium/Low/Informational]
Recommendations:
  1. [Immediate remediation step]
  2. [Long-term hardening measure]
  3. [Monitoring/detection improvement]
```


## πŸ“š Shared Resources
> For cross-cutting methodology applicable to all vulnerability classes, see:
> - [`_shared/references/elite-chaining-strategy.md`](../_shared/references/elite-chaining-strategy.md) β€” Exploit chaining methodology and high-payout chain patterns
> - [`_shared/references/elite-report-writing.md`](../_shared/references/elite-report-writing.md) β€” HackerOne-optimized report writing, CWE quick reference
> - [`_shared/references/real-world-bounties.md`](../_shared/references/real-world-bounties.md) β€” Verified disclosed bounties by vulnerability class

## References
- OWASP: [LLM Top 10 - Prompt Injection](https://owasp.org/www-project-machine-learning-security-top-10/)
- Anthropic: [Red Teaming Language Models](https://www.anthropic.com/index/red-teaming-language-models)
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why β€” repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" β€” or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself β€” your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall