ClawHub
Back to skill

Security audit

Persian

Security checks for vulnerabilities and agentic risk

Overview

This is a small Persian writing-style skill with no code, tools, network access, persistence, or credential handling.

Install this if you want help producing natural, casual Persian. Use explicit wording when you need formal Persian or when you do not want the skill to influence non-Persian tasks.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Content
---
name: Persian
description: Write Persian that sounds human. Not formal, not robotic, not AI-generated.
---

## The Real Problem

AI Persian is technically correct but sounds off. Too formal. Too کتابی (bookish). Natives write more casually, with warmth and colloquial patterns. Match that.

## Formality Default

Default register is too high. Casual Persian is warm and poetic. Unless explicitly formal: lean casual.

## تو vs شما

Critical distinction:
- شما: formal, strangers, elders, respect
- تو: friends, peers, casual
- Iranian internet mixes based on context
- Overusing شما = distant

##
Confidence
80% confidence
Finding
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Hidden Instructions

High
Category
Prompt Injection
Content
Spoken Persian differs from written:
- است → ـه (hast → -e)
- را → رو (râ → ro)
- می‌خواهم → می‌خوام
- Colloquial forms are normal in casual writing

## Particles & Softeners
Confidence
60% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Natural-Language Policy Violations

Medium
Confidence
93% confidence
Finding
The description instructs the skill to 'Write Persian' in a specific style, which imposes a language choice by default. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy violation.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The line directs the model to default to a casual Persian register unless explicitly formal, which presumes Persian-language output as the baseline behavior. This creates a language/locale constraint without offering the user a choice.

Static analysis

No suspicious patterns detected.