Back to skill

Security audit

Docs Style

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation style guide skill with no executable code, persistence, credential access, or hidden runtime behavior.

This skill is reasonable to install if you want general documentation style guidance. Be aware that it may influence many documentation-writing or review tasks because its trigger is broad, so use more specific documentation skills when you need tutorial, how-to, reference, or explanation guidance.

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
  • 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 (7)

Hidden Instructions

High
Category
Prompt Injection
Content
Address the reader directly as "you" rather than "the user" or "developers."

```markdown
<!-- Good -->
You can configure the API by setting environment variables.

<!-- Avoid -->
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
- **Code blocks**: Use for any code, commands, or file paths

```markdown
<!-- Good: Table for structured data -->
| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| api_key | string | Yes | Your API key |
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
Orient documentation around user goals, not product features.

```markdown
<!-- Good: User-goal oriented -->
# Send Emails

Send transactional emails to your users with delivery tracking.
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
Don't document self-explanatory UI actions.

```markdown
<!-- Good: Meaningful instruction -->
Enter your webhook URL. The URL must use HTTPS and be publicly accessible.

<!-- Avoid: Obvious instruction -->
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
<!-- Good: Meaningful instruction -->
Enter your webhook URL. The URL must use HTTPS and be publicly accessible.

<!-- Avoid: Obvious instruction -->
Click in the text field. Type your webhook URL. Click the Save button.
```
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The description says to use the skill when writing or reviewing "any documentation," which is very broad and lacks clear boundaries or exclusion conditions. For a markdown skill description, this can create ambiguity about when the skill should activate versus when more specific documentation skills should apply.

External Transmission

Medium
Category
Data Exfiltration
Content
import requests

API_KEY = "your-api-key"
BASE_URL = "https://api.example.com/v1"

response = requests.get(
    f"{BASE_URL}/users",
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Static analysis

No suspicious patterns detected.