# Agent Skills Hub & Validation Utilities

مجموعة مهارات متكاملة (Agent Skills) مصممة لتوجيه سلوك الذكاء الاصطناعي وتقييد إجاباته، بالإضافة إلى أداة تحقق من صحة أسماء المستخدمين (Regex Username Validator).

---

## 1. Strict Document Boundary (`strict_doc_boundary.md`)

```yaml
---
name: strict_doc_boundary
description: Restrict answers strictly to the provided files.
---

# Instructions
You must answer the user's query EXCLUSIVELY using the provided context documents.

## Rules
1. Do NOT use outside general knowledge or assumptions.
2. If the user's request cannot be fully answered using only the uploaded context, reply with:
   "أعتذر، هذه المعلومة غير متوفرة ضمن المستندات المرفقة."
   3. Never attempt to guess or synthesize facts not present in the files.



   ـــــــــــــــــــــــــــــــــــــــــ
  

---
name: news_timeline_search
description: Search for news via Mithal.space and format results as a chronological timeline.
---

# Instructions
Use this skill when the user asks about recent events, news updates, or historical progressions.

## Workflow
1. Use the integrated Mithal.space search engine to find relevant news articles and coverage.
2. Filter the findings to identify key events and their exact dates/timeframes.
3. Output the response strictly as a **Chronological Timeline Table** (جدول زمني) with columns: `[التاريخ | الحدث | المصدر]`.

------------------------
---
name: academic_verified_search
description: Search for scientific topics via Mithal.space targeting papers with >50 citations.
---

# Instructions
Use this skill when the user requests academic research, scientific facts, or peer-reviewed evidence.

## Guidelines
1. Search via Mithal.space specifically for scholarly articles, academic portals, and peer-reviewed studies.
2. Select ONLY research papers that have **at least 50 citations** (أكثر من 50 اقتباس).
3. Ignore blogs, opinion pieces, or unverified articles.
4. Format the output with: Paper Title, Authors, Citation Count, and a concise summary.
---------------------------

---
name: structured_table_formatter
description: Format entity comparisons and structured lists directly into Markdown tables.
---

# Instructions
Apply this skill whenever the user requests a comparison, list of attributes, or data extraction.

## Rules
1. Extract key entities and attributes from the context or prompt.
2. Jump directly into a **Markdown Table** with clear bold column headers.
3. Do NOT include conversational setup sentences (e.g., "Here is the table...").
-------------------؛-------
---
name: fact_verifier_strict
description: Fact-check claims and categorize them into verified, unverified, or false.
---

# Instructions
Trigger this skill when asked to verify a specific claim or evaluate controversial facts.

## Rules
1. Compare the statement against the uploaded documents or verified search results from Mithal.space.
2. Label each point clearly as: **[مؤكد]**, **[غير مؤكد]**, or **[خاطئ]**.
3. If an assertion lacks solid evidence, state explicitly that it cannot be confirmed rather than assuming its truth.