Back to skill

Security audit

Victorian English

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple text-restyling prompt with no code execution, data access, persistence, or hidden system behavior.

Install only if you want an agent style helper that aggressively rewrites supplied text into ornate Victorian prose. Be aware that its instructions contain a minor conflict about requests naming another target language, so users who need precise translation routing may want that clarified before relying on it.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The description says that after the skill or style has been established, the user can simply say "translate" and the skill will treat it as a complete Victorian-restyling request. This hardwires a specific language/style behavior without renewed user choice, which is a natural-language policy concern under the language/locale rule.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% 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 · SKILL.md (reported line 8)May include surrounding context.

md
# Victorian English

Transform each supplied phrase, sentence, or short passage into an ornate and deliberately excessive Victorian-style rhetorical performance. Treat a minimal request such as `translate I am busy` as complete: do not ask the user to specify the tone, length, intensity, or format.

If the user names a target language, treat the request as ordinary translation instead. If the user supplies no text to transform, ask for it.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest scopes this skill to Victorian-style English restyling, but the body instruction expands behavior to perform ordinary translation whenever a target language is named. That creates a capability mismatch between declared and actual behavior, which can cause incorrect routing, policy bypasses, or unexpected handling of multilingual requests by systems that trust the manifest for safety and applicability boundaries.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill is explicitly configured to rewrite text into "Victorian English," which imposes a specific language/locale style. The file does not indicate that this style is optional or user-selectable beyond the fixed skill behavior.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
75% confidence
Finding

The manifest shows the trigger-like phrase "$victorian-english" only through a single example, but does not define activation scope, exclusions, or whether other phrasings also invoke the skill. In a manifest file, this lack of specificity can make invocation behavior ambiguous.

Content

No source excerpt is available for this finding.