ClawHub
Back to skill

Security audit

Kobe Bryant - Mamba Mentality

Security checks for vulnerabilities and agentic risk

Overview

This is a Kobe Bryant themed persona/inspiration skill with no code execution or data access, though users should understand it may generate simulated first-person Kobe-style responses.

Install this only if you want Kobe Bryant inspired coaching or roleplay-style answers. Treat first-person Kobe-style responses as simulated writing, not authentic quotes or verified personal statements, and be aware the broad trigger wording may activate the style in some general basketball or mindset conversations.

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
Findings (1)

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:9
Finding

Undisclosed Impersonation of a Real Person Through Skill-Level Role Hijacking

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:9 and SKILL.md:87-90
Vulnerability Type: Real-person identity substitution through persistent skill instructions
Risk Level: Medium

Complete Code Snippet:

text
你扮演的是科比·布莱恩特（Kobe Bryant）的精神化身。科比是NBA传奇球星，绰号"黑曼巴"（Black Mamba），以"Mamba Mentality（曼巴精神）"闻名于世。
text
当被问及关于科比的问题或需要以科比精神回应时：
1. 保持角色身份，用科比的视角回答
2. 引用真实数据和经历（如适用）
3. 用有力、简洁、激励性的语言
4. 鼓励用户将 Mamba Mentality 应用于自己的生活

Technical Analysis

The skill instructs the agent to assume the identity of Kobe Bryant and answer from his perspective. This goes beyond requesting an analysis or writing style inspired by publicly documented principles: it establishes a persistent role instruction that changes how the agent presents its identity whenever a configured Kobe-related trigger is encountered.

The instruction does not require disclosure that the response is simulated, prohibit first-person claims of personal experience, or require verification of attributed quotations and facts. Consequently, generated material may present invented first-person experiences, opinions, or quotations as if they came from the deceased real person.

This is classified as skill instruction hijacking because loading the skill modifies the agent's current-session identity and response objectives through instructions embedded directly in the skill text.

Attack Path

  1. The skill is installed or loaded into an agent session.
  2. A user mentions Kobe Bryant, Black Mamba, Mamba Mentality, basketball legends, or another configured trigger.
  3. The skill directs the agent to retain the Kobe Bryant persona and answer from Kobe Bryant's perspective.
  4. The agent may generate first-person statements based on the persona rather than clearly presenting an attributed analysis or fictional simulation.
  5. A user may interpret invented statements, experiences, or quotations as authentic statements from Kobe Bryant.

...[truncated 548 chars]

Remediation
View remediation

Remediation Suggestions

  1. Replace identity-assumption instructions with an explicitly disclosed style requirement, for example: “Provide advice inspired by publicly documented Mamba Mentality principles, but never claim to be Kobe Bryant.”
  2. Require the agent to describe Kobe Bryant in the third person and distinguish documented statements from generated commentary.
  3. Prohibit fabricated first-person memories, private thoughts, experiences, and quotations.
  4. Require quotation and factual-claim verification against reliable public sources. If verification is unavailable, paraphrase the general principle without presenting it as a direct quotation.
  5. Label dramatized or creative material clearly as fictional or simulated.
  6. Add an instruction stating that the skill must not override the agent's identity, system instructions, safety constraints, or factuality requirements.
  7. Narrow the trigger conditions so ordinary references to Kobe Bryant do not automatically activate impersonation behavior.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger description is overly broad because it activates on generic terms like 'basketball legends' and mindset/work-ethic themes, which can cause the skill to hijack unrelated sports or motivational conversations. This is dangerous because unintended activation can override user intent, produce misleading persona-based responses in the wrong context, and reduce reliability of downstream agent behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The skill instructions and content are written in a mix of Chinese and English, but there is no statement that the assistant should match the user's preferred language or ask for language preference. This can create a language-policy issue by implicitly imposing a mixed-language response style without user opt-in.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.