Back to skill

Security audit

World Building

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed creative-writing helper for building and updating Vietnamese worldbuilding lorebooks, with no evidence of hidden execution, data access, or persistence outside its generated document.

Install this if you want a Vietnamese-oriented lorebook/worldbuilding workflow. Expect long structured Markdown outputs and user-facing lorebook updates; review generated canon and hooks because the skill can add creative details, but it does not show security-sensitive behavior.

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
  • 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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The description explicitly states the skill is for 'Vietnamese text‑based experiences' and ties it to the 'viết dàn ý' task, which imposes a language/locale constraint in natural language. The file does not offer an opt-in choice for other languages or explain a compliance or region-specific reason for requiring Vietnamese.

Content

No source excerpt is available for this finding.

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 · SKILL.md (reported line 410)May include surrounding context.

md
1. **Prose Slip** — You begin writing like a story: characters, actions, emotions. STOP. Rewrite in encyclopedia style.
2. **POV Creep** — You add “He felt…”, “She saw…”. STOP. Lorebook has no POV.
3. **Sensory Creep** — You describe “sunlight piercing the clouds”. STOP. Remove, unnecessary.
4. **Premise Skipped (Build Mode)** — You generate without asking the premise first. STOP. Ask 3–5 questions.
5. **No Cross‑Ref** — An entry stands alone, no link to others. STOP. Add.
6. **No Hook** — Bible complete with <12–20 open hooks. STOP. Inject more.
7. **Generic Power System** — “Cultivation has 9 realms” with no concrete limits, no cost, no “impossible”. STOP. Each realm must have a clear **price**.

Static analysis

No suspicious patterns detected.