ClawHub
Back to skill

Security audit

Radarr

Security checks across malware telemetry and agentic risk

Overview

This Radarr skill is mostly a normal media-library integration, but it needs Review because it can delete media files and can turn on future automatic collection additions without a clear opt-in step.

Install only if you are comfortable giving this skill control over your Radarr library. Confirm the exact movie before using remove, avoid --delete-files unless you intend to delete media from disk, and treat add-collection as potentially enabling future automatic collection additions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill documents shell execution (`bash scripts/radarr.sh ...`) but does not declare corresponding permissions or capabilities in a transparent way. This creates a trust and review gap: users or orchestration systems may treat the skill as lower risk than it really is, even though it can invoke local commands and interact with a privileged media-management API.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The declared purpose says the skill searches for and adds movies, but the documented behavior also includes removal of movies, optional deletion of files, existence checks, configuration enumeration, and collection monitoring. This mismatch is dangerous because it can mislead users and automated policy systems about the skill's real authority, including destructive actions that can delete media files.

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The manifest/description omits the movie removal capability even though the skill documentation includes `remove` commands and an option to delete files. Hidden destructive functionality increases the chance of accidental or unauthorized use because reviewers and users may only expect add/search behavior from the declared description.

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
The skill metadata says it searches and adds movies, but the script also supports removing movies and optionally deleting their files. This hidden destructive capability increases the chance that a caller or higher-level agent invokes deletion without informed user consent, causing loss of library entries or media files.

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The add-collection flow does more than add current movies: it also enables collection monitoring and searchOnAdd, which changes ongoing behavior and can auto-add future releases. Undisclosed persistent state changes are risky because users may believe they are performing a one-time add while actually authorizing continued automated actions.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script exposes destructive deletion behavior through --delete-files, but the usage/help output does not clearly warn that media files can be permanently removed. In an agent setting, understated destructive options can lead to accidental irreversible data loss if the command is selected from incomplete help text.

External Transmission

Medium
Category
Data Exfiltration
Content
fullCollection=$(curl -s -H "$AUTH" "$API/collection/$collectionId")
      updatePayload=$(echo "$fullCollection" | jq '. + {monitored: true, searchOnAdd: true}')
      
      updateResult=$(curl -s -X PUT -H "$AUTH" -H "Content-Type: application/json" -d "$updatePayload" "$API/collection/$collectionId")
      
      if echo "$updateResult" | jq -e '.monitored' > /dev/null 2>&1; then
        echo "👁️ Collection monitored (new releases auto-added)"
Confidence
88% confidence
Finding
curl -s -X PUT -H "$AUTH" -H "Content-Type: application/json" -d

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.