Back to skill

Security audit

Meme & Culture

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent brand meme-advice guide with explicit human review and safety limits, and it does not install code or request hidden authority.

This is reasonable to install as an advisory writing skill. Users should still keep human review before publishing brand content, especially where copyrighted material, likeness rights, sensitive events, or account posting permissions are involved.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (2)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/formats-fit-and-safety.md (reported line 25)May include surrounding context.

Safety -- the "don't" list

text
NO tragedies / disasters / sensitive events    NO divisive politics / culture-war bait
NO private moments / real people without consent NO appropriation of a culture/subculture that isn't yours
NO punching down at any group                    RESEARCH the origin + connotations first   ("if you have to ask, don't")

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

This markdown file describes 'The go / no-go a request should clear' but never specifies what kinds of user requests invoke this framework or what requests are out of scope. Without explicit trigger phrases, boundaries, or negative examples, the activation condition is ambiguous and could cause the skill to be applied to general content requests beyond meme/culture participation decisions.

Content

No source excerpt is available for this finding.