Security audit

Brainstorm

Security checks for vulnerabilities and agentic risk

Overview

This is a simple brainstorming prompt skill with no code, network use, credential access, persistence, or destructive behavior.

This skill is appropriate if you want the agent to respond more readily with ideas and creative options. Be aware it may activate for broad requests like asking for options or help thinking something through, but the inspected package does not contain code or privileged behavior.

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

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The activation examples are very broad, everyday phrases such as requests for options or help thinking through something. In an agent environment, this can cause the skill to trigger in many normal conversations where the user did not explicitly ask for brainstorming, creating unintended routing and increasing the chance of context-inappropriate behavior.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The phrase "Detect from context" creates an underspecified activation rule, allowing the skill to engage based on broad or implicit cues rather than an explicit user request. In an agent system, this can cause unintended mode switching or unsolicited behavior, which increases the chance of the skill influencing responses when the user did not clearly ask for brainstorming assistance.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The phrase "Use when stuck or need more variety" does not clearly define when this skill should activate or what constraints limit its use. It overlaps with common conversational situations and provides no negative examples or specific trigger scope, which could lead to unintended invocation.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The instruction "When user seems stuck, offer:" relies on a subjective and ambiguous condition rather than explicit, observable triggers. Without clearer criteria or exclusions, different implementations may invoke the skill inconsistently or too broadly.

Static analysis

No suspicious patterns detected.