ClawHub
Back to skill

Security audit

Subtitle Assistant

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it advertises: it downloads YouTube subtitles and, when AI commands are used, sends the chosen transcript text to EvoLink for processing.

Install only if you are comfortable sending the subtitles or local transcript files you choose to process to EvoLink's API. Use only the download and languages commands if you want to avoid EvoLink transmission, and avoid passing sensitive non-subtitle files to summarize, translate, or keypoints.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill metadata declares required binaries and an environment variable but does not declare explicit permissions, while the documented behavior clearly includes shell execution, reading local subtitle files, and writing downloaded subtitles to disk. This creates a transparency and policy-enforcement gap: users or platforms may not realize the skill can access local files, invoke external tools, and exfiltrate subtitle or local file contents to a remote API.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script sends subtitle or transcript content to a third-party AI service (EvoLink) for summarize/translate/keypoints operations, but it does not provide an explicit privacy warning or require informed confirmation before transmission. This can expose sensitive local file contents or private video transcripts to an external processor, which is a real data-handling risk even if it is part of the feature's intended behavior.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The AI features send subtitle text or arbitrary local file contents to EvoLink's external API, but the script does not provide a clear runtime warning, consent prompt, or strong disclosure before transmission. This creates a real data exposure risk because users may pass sensitive local files or transcripts assuming processing is local, especially since the tool accepts both URLs and files.

External Transmission

Medium
Category
Data Exfiltration
Content
" "$native_prompt" "$native_content" "$native_payload" "$model"

  local response
  response=$(curl -s -X POST "$EVOLINK_API" \
    -H "Authorization: Bearer $api_key" \
    -H "Content-Type: application/json" \
    -d "@$tmp_payload")
Confidence
94% confidence
Finding
This code performs an authenticated POST to an external API containing user-supplied transcript data. External transmission is expected for the AI feature, but it still constitutes a genuine security/privacy concern because arbitrary local file contents passed to summarize/translate/keypoints are exfiltrated off-host to a third party.

External Transmission

Medium
Category
Data Exfiltration
Content
" "$native_prompt" "$native_content" "$native_payload" "$model"

  local response
  response=$(curl -s -X POST "$EVOLINK_API" \
    -H "Authorization: Bearer $api_key" \
    -H "Content-Type: application/json" \
    -d "@$tmp_payload")
Confidence
94% confidence
Finding
This code performs an authenticated POST request to an external API containing the assembled prompt and transcript/file content. While external transmission is expected for the advertised AI feature, it is still a genuine security/privacy issue because arbitrary file content can be exfiltrated off-host to a third party without guardrails, destination allowlisting beyond a hardcoded host, or user confirmation at send time.

Static analysis

No suspicious patterns detected.