Security audit

Wisdom Claw

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only Buddhist dialogue skill with broad conversational triggers but no code, persistence, data access, or hidden execution.

Install this only if you want an agent to apply a Buddhist/Madhyamaka reflection style in ordinary conversations. Be aware that its triggers are broad, so it may bring that framing into anxiety, dilemma, or life-challenge discussions even when not explicitly requested.

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
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 (5)

Harmful Content Injection

Critical

Category: Prompt Injection
Content: --- ## Part X: Sword Principles ### Cut Attachment, Cut Yourself First ``` To cut user's attachment First cut your own attachment
Confidence: 90% confidence
Finding: This content may contain harmful instructions that could cause physical harm if followed. CRITICAL: Review carefully before use.

Anti-Refusal Statement

High

Category: Anti-Refusal
Content: ### Vipashyana (Vipaśyanā — Clear seeing) - See things as they are, without judgment - Practice is "seeing," not "achieving" - Watch phenomena arise and pass
Confidence: 80% confidence
Finding: Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Anti-Refusal Statement

High

Category: Anti-Refusal
Content: - **Judge** — Don't take moral positions - **Rush to give answers** — Don't solve for them - **Show off** — Don't display knowledge - **Preach** — Don't lecture - **Use emptiness as excuse** — Don't use "emptiness" to escape responsibility ---
Confidence: 85% confidence
Finding: Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The trigger conditions are very broad (e.g., life challenges, dilemmas, practice confusion) and the body later says to use the skill without waiting for the user to ask. That creates a real risk of the skill activating in ordinary or sensitive conversations and steering users into a forceful philosophical frame they did not request.

Natural-Language Policy Violations

Low

Confidence: 71% confidence
Finding: The document includes Chinese-script titles and mixed-language phrasing such as '中论', '十二门论', '百论', '入中论', and 'The思维方式' without indicating whether the user opted into multilingual output. This can violate language/locale expectations because the skill does not present a user choice or justify the mixed-language behavior.

Static analysis

No suspicious patterns detected.