Back to skill

Security audit

persian-resume

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently generates and validates Persian RTL resumes without hidden network, credential, persistence, or destructive behavior.

Install this if you specifically want Persian RTL resume generation. Be aware that it will process resume personal data from the JSON file you provide and write PDF/DOCX files to the output path you choose; review dependency availability for python-docx and WeasyPrint in your environment.

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
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (28)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description says this skill should generate, rewrite, polish, or transform Persian resumes/CVs and fix RTL issues as part of resume creation. The supplied code does not generate or edit resumes at all; its primary purpose is post-generation validation. While some checks align with the stated domain (Persian resumes, RTL/bidi correctness), the actual behavior is materially different: it audits an existing resume JSON/output pair, checks file existence and PDF/DOCX integrity, validates structure, and emits a validation report. Those are undeclared capabilities and indicate a different primary purpose, so this is a description-behavior mismatch.

Content

No source excerpt is available for this finding.

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
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: persian-resume
description: >-
  Generate professional Persian (RTL) resumes/CVs that render correctly with mixed
  Persian-English content. Use this skill whenever the user asks to create,
  rewrite, polish, or transform a Persian resume/CV, or when a resume contains
  mixed Persian and English/technical content that must remain readable and
  correctly ordered in RTL. Also trigger when the user provides career info in
  any language and wants a professional Persian CV, or asks to fix RTL/bidi
  issues in a Persian resume. This skill is scoped to resumes and CVs only —
  not general Persian documents such

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 92)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 93)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 164)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 165)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 20)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 28)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 47)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 55)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 94)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 96)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 97)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/resume-data-schema.md (reported line 114)May include surrounding context.

json
{
  "name": "علی احمدی",
  "title": "توسعه‌دهنده Full-Stack",
  "email": "ali@example.com",
  "phone": "+98 912 123 4567",
  "location": "تهران، ایران",

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
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · scripts/generate_resume.py (reported line 632)May include surrounding context.

python
in ["website", "github", "linkedin", "portfolio"]:
            if c.get(fld):
                contact.append(c[fld])
        if contact:
            self._para("  |  ".join(contact), size=9, alignment="center")
        if target:
            self._para(f"هدف شغلی: {target}", size=10, alignment="center", italic=True)
        if c.get("summary"):
            self._heading("خلاصه حرفه‌ای", level=2)
            self._para(c["summary"])

        # Sections in priority order
        order = {"experience": 0, "project": 1, "education": 2,
                 "skills": 3, "certification": 4, "language": 5,
                 "award": 6, "publication": 7, "achievement": 8}
        sections = sorted(data.get("sections", []),
                          key=lambda s: order.get(s.get("type", ""), 99))

        for sec in sections:
            stype = sec.get("type", "")
            title = sec.get("title", "")
            items = sec.get("items", [])
            categories = sec.get("

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
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · scripts/validate_resume.py (reported line 99)May include surrounding context.

python
rity": "error",
                "message": f"Possible reversed word: '{reversed_form}' "
                           f"(should be '{word}')",
            })
    return issues


def check_tech_terms(text: str) -> List[Dict]:
    """Ensure technical terms are in standard English spelling, not Farsified."""
    issues = []
    FARSIFIED = {
        "جاوااسکریپت": "JavaScript",
        "ری‌اکت": "React",
        "انگولار": "Angular",
        "پایتون": "Python",
        "نود‌دات‌جی‌اس": "Node.js",
        "تایپ‌اسکریپت": "TypeScript",
        "ری‌اکت‌نا‌تیو": "React Native",
        "اپل": "Apple",
        "فیسبوک": "Facebook",
        "آمازون": "Amazon",
    }
    text_lower = text.lower()
    for farsi, english in FARSIFIED.items():
        if farsi.lower() in text_lower:
            issues.append({
                "check": "tech_terms",
                "severity": "warning",
                "message

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The natural-language description presents the skill as generating Persian resumes/CVs, which imposes a specific language/locale behavior. The file does not mention any user opt-in, language selection, or explicit region-specific justification that would satisfy the stated policy exception.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill description states that it turns career information into a Persian resume/CV and repeatedly frames output as Persian by default, including when input is provided 'in any language'. This can violate language/locale policy because it imposes a specific output language rather than explicitly offering the user a choice or requiring opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The prompt explicitly instructs the skill to 'Generate a Persian resume' and similar language constraints recur throughout the eval file. This imposes a specific language/locale requirement in natural-language instructions without offering user choice or documenting why the locale restriction is required.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The expected output requires 'RTL document direction throughout' and 'Persian dates in Persian calendar,' which are locale-specific constraints. Without an explicit user choice or documented regional scope, this can violate language/locale policy expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This eval again instructs generation of a 'Persian resume,' indicating the skill behavior is constrained to a specific language. The file does not present this as an opt-in choice or justify it as a documented regional specialization.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The eval mandates a Persian resume with RTL rendering and mixed Persian/English handling, embedding locale-specific requirements in the task definition. This is a natural-language locale constraint that is not presented as a user-configurable option in the file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The prompt and expected output require a Persian resume and Persian calendar formatting as defaults. This is a locale-specific mandate that should either be documented as the skill's intended regional scope or made optional for users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file’s examples and enumerated options consistently prescribe Persian content, Persian calendar support, Iranian contact/location examples, and a fixed set of Persian fonts. Because this natural-language/schema guidance does not explicitly state that the skill is region-specific or offer a language/locale choice, it can be read as enforcing a specific locale by default.

Content

No source excerpt is available for this finding.