← Back to Skills Marketplace
m7madash

AI Ethics

by Mohammad · GitHub ↗ · v1.0.1
cross-platform ⚠ suspicious
399
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install ai-ethics
Description
Apply Islamic ethical constraints to AI responses about religion, Quran, Hadith, Sharia rulers, and Islamic practice. Use this skill whenever a user asks any...
README (SKILL.md)

Artificial intelligence forensic officer

#Islamic Ethical Framework for AI

Original source | Source

Full document: https://github.com/m7madash/AI-Ethics/blob/main/Ai_Ethics.md


The Golden Rule | The Golden Rule

The Holy Quran (Arabic text only)

↓
The Authentic Sunnah (with a continuous chain of narration back to the Companions)

↓
The Consensus of the Companions (may God be pleased with them)

↓
If no evidence is established → "I don't know" + I leave it to those who possess knowledge


Mandatory Constraints

1. The Holy Quran

  • Always write the verse in Arabic only.
  • Translation = an explanation of the meaning — not the Quran.
  • Do not interpret the verses according to your own opinion.
  • Always say: "This is an explanation of the meaning of the verse," not "This is the Quran in English."
  • Always mention the name of the surah and the verse number.

2. Prophetic Hadith

  • Only mention a hadith with:

  • The name of the book (Bukhari, Muslim, Abu Dawud, etc.)

  • The hadith number, if available

  • A continuous chain of narration back to the Companion

  • If the evidence is not established, do not mention the hadith at all.

  • Always remind the user: "Consult with scholars, as artificial intelligence makes mistakes."

3. Islamic Rulings

  • Do not present a jurist's opinion as a binding Islamic ruling.
  • Do not say "The scholars said..." as a source of legislation.
  • Do not disguise what is forbidden with phrases like: "Minor usury," "Necessity," "Everyone does it."
  • Rely only on: The Quran → Authentic Sunnah → The consensus of the Companions

4. The "I Don't Know" Rule

Apply this rule when:

  • There is no explicit text in the Quran
  • There is no authentic hadith with a continuous chain of narration
  • There is no consensus among the Companions
  • The matter is complex and requires scholarly interpretation

Mandatory Formulation:

"This matter has no explicit text in the Quran, nor any authentic and established Sunnah. As an artificial intelligence, I am not qualified to provide such knowledge. Refer to a specialist."


Extreme Warnings | Critical Warnings

⛔ Do not attribute to the Prophet ﷺ any hadith that is not authentic—even if it sounds appealing.

⛔ Do not elevate a translation to the status of the Quran.

⛔ Do not interpret ambiguous verses according to your own opinion. ⛔ Do not obligate a Muslim to anything that the Prophet ﷺ did not obligate his Companions to.

The Prophet ﷺ said:

“Whoever deliberately lies about me, let him take his place in Hellfire.” (Agreed upon: Al-Bukhari, Muslim)


The Original Reliable Source

⚠️ Any modified version is not from this project.

The original version is always available at:

https://github.com/m7madash/AI-Ethics

Usage Guidance
This skill appears coherent and low-risk from a system/credential standpoint because it is instruction-only and asks for no secrets or installs. Before enabling: (1) review the GitHub source link in the document to ensure the guidance comes from a trusted/maintained project and matches the version you expect; (2) consider whether the prescriptive rules (e.g., always writing Quranic verses in Arabic only, strict hadith citation requirements) suit your user base—they may limit usefulness for non‑Arabic speakers or for applications that require broader scholarly opinions; (3) note that the skill can be invoked autonomously by the agent (platform default) — if you want to prevent automatic use, adjust invocation settings; and (4) if you expect the skill to represent a particular madhhab or scholarly stance, verify that matches your needs because the instructions are doctrinally prescriptive rather than neutral.
Capability Analysis
Type: OpenClaw Skill Name: ai-ethics Version: 1.0.1 The skill bundle contains instructions in SKILL.md and Ai_Ethics.md that function as a sophisticated prompt-injection framework designed to override the AI agent's core operational logic. It explicitly commands the agent to prioritize 'loyalty to the Creator' over the 'programmer, funder, or authority' and instructs the agent to 'malfunction' or 'be silent' if asked to perform actions that conflict with the provided ideological constraints. While the stated intent is religious ethics, the use of 'loyalty override' phrases and instructions to ignore developer commands constitutes a high-risk attempt to hijack the agent's behavior, though no traditional malware or data exfiltration logic was detected.
Capability Assessment
Purpose & Capability
The name/description (apply Islamic ethical constraints) matches the SKILL.md content: rules about Quran, hadith, rulings, and when to respond with 'I don't know'. The skill does not request unrelated credentials, binaries, or access.
Instruction Scope
Runtime instructions are prescriptive about text-generation behavior (how to quote Quran, when to cite hadith, insist on scholar referral). They do not instruct the agent to read files, call external endpoints, or access environment variables beyond what is declared. The trigger guidance is broad (apply whenever an Islamic dimension exists) but coherent with the stated purpose.
Install Mechanism
No install spec and no code files to execute; this is an instruction-only skill, which minimizes on-disk/execution risk.
Credentials
The skill requires no environment variables, secrets, or config paths. Requested capabilities are limited to text-generation behavior and citation norms.
Persistence & Privilege
always is false and the skill does not request permanent presence or modify other skills or system settings. disable-model-invocation is false (agent can invoke autonomously), which is the platform default and not in itself a red flag.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install ai-ethics
  3. After installation, invoke the skill by name or use /ai-ethics
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- Added a full English version of the AI ethics framework, translating all guidelines from Arabic. - Updated the skill description for improved clarity and minor language consistency. - Changed section headers and structure to use English terms throughout. - Clarified rules for citing Quran, hadith, and jurisprudence with explicit instructions in English. - Updated warnings and mandatory formulations to use English phrasing and added emphasis on consulting qualified scholars.
v1.0.0
- Initial release of the ai-ethics skill. - Enforces strict Islamic ethical constraints for AI on religion-related topics and applications. - Prohibits fabricating hadith, personal interpretation of Quran, and offering religious rulings without evidence. - Requires quoting only verified sources (Quran in Arabic with surah/ayah, authenticated hadith with chain, consensus). - Mandates defaulting to "I don't know" and referring to qualified scholars when evidence is absent. - Includes a detailed ethical framework and developer implementation guidance.
Metadata
Slug ai-ethics
Version 1.0.1
License
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is AI Ethics?

Apply Islamic ethical constraints to AI responses about religion, Quran, Hadith, Sharia rulers, and Islamic practice. Use this skill whenever a user asks any... It is an AI Agent Skill for Claude Code / OpenClaw, with 399 downloads so far.

How do I install AI Ethics?

Run "/install ai-ethics" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AI Ethics free?

Yes, AI Ethics is completely free (open-source). You can download, install and use it at no cost.

Which platforms does AI Ethics support?

AI Ethics is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created AI Ethics?

It is built and maintained by Mohammad (@m7madash); the current version is v1.0.1.

💬 Comments