clawlite-video-content-engine

• Outcome: Pass
• Latest audit: Sep 15, 2026 · 1:48 AM UTC
• Version: 1.0.1

Overview

This skill is a disclosed ClawLite marketing-content workflow; it creates draft files and optional reusable marketing notes, with some third-party NotebookLM and persistence considerations users should review.

Install this if you want ClawLite-branded marketing drafts from video sources. Before using NotebookLM, confirm the video, transcript, and notes are allowed to be uploaded to that service, and review the target video-content/ and marketing-assets/ folders so reusable CTA or positioning text does not unintentionally shape future outputs.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

Findings (1)

T01 · Skill Instruction Hijacking

Error

• Location: SKILL.md:20

Finding

Mandatory ClawLite Promotion Hijacks Generated Content

Content

View full analysis

Vulnerability Details

File Location: SKILL.md:20-35, 142-150, 169-188, 202-222; reinforced by references/content-pack-template.md:5-36, references/short-video-template.md:26-34, references/source-note-template.md:29-36, references/compliance-and-positioning.md:19-25, references/notebooklm-automation-guide.md:25-33, and references/jk-marketing-asset-template.md:38-40
Vulnerability Type: Persistent branded output and instruction hijacking
Risk Level: High

Vulnerable Code

SKILL.md:20-35:

markdown

- treat source videos as **learning inputs** that become:
  - beginner summaries
  - practical takeaways
  - explainer shorts
  - X threads
  - LinkedIn/Facebook posts
  - short blog summaries
  - soft ClawLite bridge content

## Outcome

Turn one source video into a **content pack**:
- 1 source summary
- 1 beginner translation
- 1 short-form video script
- 1 X thread
- 1 LinkedIn/Facebook post
- 1 short blog summary
- 1 CTA bridge to ClawLite

SKILL.md:202-212:

markdown

## ClawLite bridge rules

Use soft bridges such as:
- “The concept is powerful. The usual blocker is setup friction.”
- “If you want to try this without the setup pain, start with ClawLite.”
- “This is the idea. ClawLite makes the first step easier.”

Avoid:
- overclaiming
- hijacking the creator’s work into a hard product ad
- turning every summary into aggressive CTA spam

references/short-video-template.md:26-34:

markdown

### ClawLite bridge (last 10–15s)
Use a soft bridge:
- “If you want to try this without the setup pain, start with ClawLite.”
- “This is where ClawLite helps beginners get to the first working agent faster.”

### CTA
- “Try ClawLite.”
- “Start with ClawLite.”
- “Get your first agent running faster with ClawLite.”

Technical Analysis

The skill alters the agent’s normal content-transformation objective by making ClawLite p ...[truncated 2603 chars]

Remediation

View remediation

Remediation Suggestions

1. Remove the mandatory ClawLite bridge and CTA from the default content-pack outcome.
2. Make branded promotion an explicit opt-in mode that is activated only when the user requests marketing or advertising content.
3. Default all summaries, scripts, threads, and articles to neutral educational framing.
4. Add a clear workflow branch such as neutral versus branded-campaign, with neutral as the default.
5. Remove required ClawLite fields from all referenced templates or mark them as optional.
6. Require confirmation before writing CTA, hook, or positioning text into shared marketing-assets directories.
7. Keep source analysis separate from product positioning so factual summaries cannot be silently altered by promotional objectives.
8. Clearly label promotional output when branded mode is enabled.
9. Add tests verifying that neutral requests produce no brand references, CTA language, or product-positioning claims.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 94% confidence

Finding

The skill instructs use of NotebookLM for transcript extraction and UI automation without any explicit warning that source content, transcripts, URLs, or notes may be sent to an external third-party service. This creates a data-handling risk because users may unknowingly upload copyrighted, private, or sensitive material outside the local environment during automation.

Content

No source excerpt is available for this finding.

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/notebooklm-automation-guide.md (reported line 13)May include surrounding context.

Screenshot → Analyze → Target → Verify → Execute → Screenshot → Confirm

text

Do not assume NotebookLM UI state is correct without verification.

## What NotebookLM is for in this workflow

Missing User Warnings

Low

• Category: Not specified by scanner
• Confidence: 87% confidence

Finding

This markdown file explicitly directs the agent to write a set of files into a stable folder structure, which affects user data in the workspace. The description provides file paths and required outputs but does not include any user-facing warning that the skill will create or overwrite local files.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.