ClawHub
Back to skill

Security audit

@gerandica

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent Instagram carousel generator with some language and trigger-scope quirks, but no evidence of hidden, destructive, persistent, or credential-related behavior.

Install this if you want an agent to help produce data-backed Instagram carousel HTML. Be explicit about the desired language, brand details, topic, and whether web research is allowed; review generated statistics and sources before publishing.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Hidden Instructions

High
Category
Prompt Injection
Confidence
70% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/components.md (reported line 12)May include surrounding context.

html
<div class="slide" style="background:{SLIDE_BG};">
  <div style="position:absolute;top:28px;left:36px;right:36px;bottom:52px;min-height:0;display:flex;flex-direction:column;overflow:hidden;">
    <!-- ALL visible content goes here -->
  </div>
  <!-- Progress bar (position:absolute, sits outside the content flow) -->
  <!-- Swipe arrow (position:absolute, sits outside the content flow) -->

Hidden Instructions

High
Category
Prompt Injection
Confidence
70% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/components.md (reported line 285)May include surrounding context.

html
<!-- Note: this goes INSIDE the slide container's inner div (which already has overflow:hidden and padding-bottom:52px) -->
<div style="flex:1;display:flex;flex-direction:column;position:relative;">
  <!-- Decorative background number -->
  <div style="position:absolute;top:-8px;right:-8px;font-family:'{HEADING_FONT}';font-size:140px;font-weight:800;color:{BRAND_PRIMARY};opacity:0.08;line-height:1;pointer-events:none;user-select:none;">
    {NN}
  </div>

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README states that output language must be 'simple Russian' without indicating that this depends on user preference or locale. This can override or conflict with the user's requested language, causing unauthorized behavior, degraded usability, or policy issues in multilingual environments where respecting user intent is required.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger criteria are broad enough to activate on many generic social-media design requests, not just clearly scoped Instagram carousel tasks. Over-broad activation can cause the agent to invoke the wrong skill, collect unnecessary brand details, and steer users into a constrained workflow that may not match their request, increasing the chance of unintended data handling or misleading output.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

Forcing Russian output by default without user consent can override user expectations and produce content in the wrong language, which is especially risky for customer-facing publishing workflows. In a skill that generates ready-to-post social content, this can cause brand, compliance, or operational errors if the user intended another language but the system silently defaults to Russian.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This markdown file contains example chip labels in Russian and additional Russian trigger phrases for their use, but nowhere indicates that the skill is Russian-only or that users may choose another language. That creates a natural-language locale policy issue because the file implicitly steers output into a specific language without opt-in.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.