## https://sploitus.com/exploit?id=1257B0B3-2F17-5B09-B4A8-15A65DE8004D
# AIPOC – AI POC Thinking Framework Skill
A framework designed for AI coding assistants/safety assistants, derived from 7468 real POC examples. This framework enables AI to quickly understand POC concepts, master mapping syntax, and efficiently construct exploitation chains.
## What is this?
AIPOC is a **POC thinking skill** designed for AI coding assistants/safety assistants. It extracts core thought patterns from 7468 real vulnerability POCs in [yangyu91/poc-lib](https://github.com/yangyu91/poc-lib), enabling AI to:
1. **Understand concepts instantly**: Complete the thought process of “fingerprint → mapping → POC matching → exploitation chain” within 5 seconds when given target products/CVEs/fingerprints.
2. **Master mapping syntax**: Proficiency in constructing FOFA, Hunter, Quake, and Google Dork syntax for information gathering.
3. **Adapt POC templates**: Quickly adapt specific exploitation code from general templates.
4. **Construct chained attacks**: Combine multiple vulnerabilities into an attack chain (e.g., authentication bypass → SQLi → file upload → RCE).
5. **Verify origins**: Each concept is marked with the source POC library location, making it easier for AI and humans to refer to original POCs for verification.
## Data Source
| Source | Number of Entries | Description |
|-------|------------------|--------------|
| poc.db (POC management platform) | 4393 | Vulnerability-Wiki-POC, Sec-Fork/POC-20250106, Awesome-POC, etc. |
| Goby POC library | 3656 | POC in JSON format from the goby-poc repository |
| **Total after deduplication** | **7468** | Export date: 2026-08-10 |
## Directory Structure
```
aipoc/
├── SKILL.md # Main skill file (activation protocol, core thinking framework, decision tree)
├── README.md # This file
├── references/
│ ├── vuln-patterns.md # Deep thinking patterns for 17 types of vulnerabilities (804 lines)
│ ├── recon-syntax.md # Guide for constructing FOFA, Hunter, Quake, and Google mapping syntax (821 lines)
│ ├── poc-templates.md # 12 reusable POC request templates (1851 lines)
│ ├── source-index.md # Quick index of product name → POC library ID (370 lines)
│ └── attack-chains.md # Combination of 10 common attack chains (1160 lines)
└── scripts/
└── fofa_builder.py # FOFA, Hunter, Quake query constructor (506 lines)
```
## Coverage of Vulnerability Types
| Priority | Type | Count | Characteristics |
|--------|------|------|------|
| P0 | RCE | 1860 | Direct shell acquisition, highest value |
| P0 | Deserialization | 139 | Often unauthorized Java RCE |
| P1 | SQL Injection | 1630 | Data extraction, can escalate to RCE |
| P1 | File Upload | 698 | Webshell deployment |
| P1 | Command Injection | 61 | Direct OS command execution |
| P2 | File Reading | 1173 | Configuration/credential theft |
| P2 | Authentication Bypass | 141 | Access without credentials |
| P2 | Permission Bypass | 47 | Elevation to administrator privileges |
| P3 | Unauthorized Access | 234 | Exposure of sensitive endpoints |
| P3 | Information Leakage | 488 | Credential/key leakage |
| P3 | SSRF | 91 | Internal network penetration |
| P3 | Weak Passwords | 196 | Default/weak credentials |
## Quick Use
### As a Skill for Trae/Codex AI Assistants
Place this directory in the AI assistant skills path, and enter `aipoc` in the dialogue to activate the POC thinking mode. ### FOFA Query Builder
```bash
# Query by product name
python3 scripts/fofa_builder.py --product "Apache OFBiz"
# Output: app="Apache_OFBiz"
# Combine multiple signals
python3 scripts/fofa_builder.py --title "Backend" --body "login" --port 8080
# Output: title="Backend" && body="login" && port="8080"
# Convert to Eagle chart syntax
python3 scripts/fofa_builder.py --product HIKVISION --engine hunter
# Output: app.name="HIKVISION"
# Query known product mappings
python3 scripts/fofa_builder.py --lookup OA
# Output: Found 6 matches: Zhiyuan OA / Jinhe OA / Tongda OA / Lanling OA / Wanhuo OA / Huatian Dynamics OA
```
### Example of AI Thinking Process
```
User: "What are the vulnerabilities in the HIKVISION integrated security management platform?"
AI Thinking:
1. FOFA construction: body="hikvision" || app="HIKVISION"
2. POC library search: Found over 15 entries (IDs: 116, 1692-1706, 4680-4688)
3. Type analysis: RCE (multiple), Fastjson deserialization, file upload
4. Priority sorting: RCE entries prioritized (IDs: 1698, 1701, 1705 – front-end command execution)
5. Source library trace: https://github.com/yangyu91/poc-lib/blob/main/by_type/RCE.md
```
## Important Notes
### Source POC Library Trace
All thought processes in this skill are derived from [yangyu91/poc-lib](https://github.com/yangyu91/poc-lib). Due to possible misunderstandings or simplifications during the derivation process, please always verify the source POC library before use:
1. Search for product names or CVEs in [INDEX.md](https://github.com/yangyu91/poc-lib/blob/main/INDEX.md) to obtain POC IDs.
2. Open the corresponding `by_type/.md` file and search for the ID or vulnerability name.
3. Verify the original POC code to ensure the conditions for exploitation and the accuracy of the payload.
### Security Statement
- For use only in **security research, vulnerability reproduction, and authorized testing**.
- Source data statement: "No exploitation code shall be executed on other systems without authorization."
- Goby POC source statement: "Does it contain backdoors and self-judgment for repetition?"
- **Please read the original POC code before use** to ensure no backdoors.
- **It is strictly prohibited** to execute any exploitation code on unauthorized targets.
## Technical Features
- **Native AI Design**: All content is written in a structured format that can be directly understood by AI; non-human-readable documents.
- **Thinking Framework Priority**: Not a list of POCs, but teaches AI “how to think” about POCs.
- **Proficiency in Mapping Syntax**: Fully covers the syntax construction of FOFA/Eagle/Quake/Google platforms.
- **Source Library Trace**: Each thought process is marked with the source POC library location, enabling quick verification.
- **Attack Chain Thinking**: Focuses not just on individual vulnerabilities, but on the combined exploitation of vulnerabilities.
## License
The content of this skill is derived from publicly available POC libraries and is intended for use in security research only. Users must comply with local laws and regulations.