ClawHub
Back to skill

Security audit

Paprika Recipe Importer

Security checks across malware telemetry and agentic risk

Overview

This skill does its stated job locally, but its instructions use raw recipe-derived filenames in shell commands and shared temporary paths, which users should review before installing.

Install only if you are comfortable with a skill that creates local recipe files and runs a local Python packaging script. Use sanitized, fixed temporary filenames or a private temp directory rather than raw recipe titles, and remove intermediate JSON files after creating the `.paprikarecipes` archive.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Vague Triggers

Medium
Confidence
79% confidence
Finding
The trigger conditions are broad enough that the skill may activate whenever a user shares recipe-like content or implies export to Paprika, even if they did not clearly request file creation. Over-broad activation can cause unintended processing of user-provided content and unintended file generation, which is a security and privacy concern in agentic systems.

Session Persistence

Medium
Category
Rogue Agent
Content
## Step 2 — Build JSON

Write the recipe as a JSON file to a temp path, e.g. `/tmp/<recipe-name>.json`.

## Step 3 — Run the Script
Confidence
91% confidence
Finding
Writing recipe data to predictable paths under /tmp can expose user content to other local processes or users on shared systems, and the recipe name may contain unsafe characters that affect file handling. Temporary files can also persist longer than intended, creating unnecessary retention of potentially sensitive user data.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.