ClawHub
Back to skill

Security audit

Model-Aware Prompt Optimizer

Security checks for vulnerabilities and agentic risk

Overview

This is a prompt-rewriting skill with static Markdown guidance and no evidence of hidden execution, data access, persistence, or destructive behavior.

This skill appears appropriate for users who want help rewriting prompts. Review any generated prompt before using it, especially if your original prompt contains sensitive business rules, private data, or tool permissions, because the skill is designed to preserve those requirements in the rewritten output.

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
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • 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)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · references/google.md (reported line 5)May include surrounding context.

md
- Be precise and direct. Define ambiguous terms and parameters.
- Use one consistent structure: Markdown headings or XML-style tags. Do not mix delimiter styles without need.
- Put critical role, behavior, constraint, and output instructions in the system instruction or near the beginning.
- For large context, put the context first and the specific task or question at the end, with a clear transition such as `Based on the information above`.
- Treat text, images, audio, and video as explicit inputs and name how each should be used.
- Specify verbosity when a detailed or conversational answer is needed; Gemini 3 models tend to answer directly.

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/xai.md (reported line 13)May include surrounding context.

md
## API advice kept outside the prompt

- `reasoning_effort` is an API control on supported Grok models and should not be encoded as verbose reasoning instructions in the prompt.
- Reasoning-model parameter support varies; do not recommend sampling or stop parameters without checking the exact model documentation.
- Prompt caching benefits from stable prefixes and a stable conversation or cache key.

## Official sources

Static analysis

No suspicious patterns detected.