ClawHub
Back to skill

Security audit

Prompt Engineering

Security checks for vulnerabilities and agentic risk

Overview

This is a prompt-engineering guidance skill made of markdown references, with no code execution, network behavior, credential handling, or persistence.

This skill appears safe to install for help drafting and improving prompts. Be aware that it may activate for broad prompt-help requests and may suggest documenting prompt versions, so users should keep any saved prompt files in an appropriate project location and avoid including secrets in examples they provide.

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
  • 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 (5)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill description claims applicability to 'any LLM task' and also triggers on vague requests for help, creating an overly broad activation scope. This can cause the skill to be invoked in contexts beyond prompt engineering, potentially overriding more appropriate domain-specific or safety-constrained skills and increasing the chance of unsafe or irrelevant guidance.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### Clarity Assessment
- Is the issue clearly distinguishable from intended behavior?
- Could a developer reproduce this without asking questions?
- Are technical terms used correctly?

### Improvements Needed
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.

Scope Creep

Low
Category
Excessive Agency
Content
**What it looks like:**
```
"Analyze this from every possible angle including but not limited to technical, business, legal, ethical, environmental, social, political, and philosophical perspectives..."
```

**Why it fails:**
Confidence
70% confidence
Finding
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Natural-Language Policy Violations

Low
Confidence
90% confidence
Finding
This markdown file includes bilingual section headings such as Japanese parenthetical labels alongside English content. Under the stated policy, forcing or introducing a specific language/locale without user opt-in can be a natural-language policy issue, and this file does not explain that the bilingual format is optional or required for a specific audience.

Natural-Language Policy Violations

Low
Confidence
88% confidence
Finding
The prompt evolution example instructs translation specifically for a business audience in Mexico and further mandates locale-specific conventions such as formal 'usted', Mexican pesos, and DD/MM/YYYY formatting. This is a natural-language locale constraint presented as a default requirement, with no indication of user choice or region-specific justification.

Static analysis

No suspicious patterns detected.