ClawHub
Back to skill

Security audit

Question Explanation

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent HTML tutorial generator, but it can silently create or overwrite local HTML files even when the user only asked for an explanation.

Review this skill before installing if you expect normal chat explanations. It should be limited to cases where the user explicitly wants an HTML tutorial file, and it should use unique filenames or ask before overwriting existing files.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

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 · SKILL.md (reported line 27)May include surrounding context.

md
## Preparation
No special dependencies or manual setup are required. Rely on the agent's multimodal and reasoning abilities.

## Output Rules
- Generate silently: do not print progress, intermediate notes, summaries, or narration while working
- Create a real file: write the complete HTML into an actual `.html` file instead of only describing it in the conversation
- Save the final deliverable as a complete HTML file and send that HTML file to the user

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill description is extremely broad and could activate for many ordinary educational requests, causing this skill to take over in situations where a narrower or safer skill would be more appropriate. Over-broad triggering increases the chance of unintended file creation and unnecessary handling of user content beyond the user's explicit request.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The 'Use this skill when' section lacks limiting constraints and effectively matches almost any question explanation task, which can cause over-invocation. In context, that matters because the skill does more than answer questions: it generates files and constrains output behavior, so accidental activation has side effects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill instructs the agent to write a real HTML file and send it to the user without warning or obtaining confirmation for filesystem modification. This creates an avoidable side effect and could overwrite local artifacts or produce files the user did not intend to create.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Line L095 requires descriptive English identifiers, and the base HTML example at L103 fixes the document language to lang="en". This imposes an English locale by default without user opt-in or a stated reason that the skill must be English-only.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The instruction to 'create or overwrite' the HTML file explicitly permits destructive modification of existing files without confirmation. Even in an educational skill, silent overwrite behavior is risky because it can destroy prior work or unrelated artifacts in the working directory.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.