lovstudio-style-clone

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The documentation describes installation via the vendor's command-line tool (npx lovstudio). This is a standard part of the author's ecosystem for managing skills.
  • [PROMPT_INJECTION]: The skill instructions define a mandatory workflow for the agent. No attempts to bypass safety filters or override system-level instructions were found.
  • [COMMAND_EXECUTION]: The workflow relies on standard agent interaction tools (like AskUserQuestion) and does not involve the execution of arbitrary shell commands or untrusted scripts.
  • [DATA_EXFILTRATION]: No network exfiltration or unauthorized access to sensitive file paths (such as credentials or SSH keys) was detected. The skill writes style profiles to a specific local directory (./output/style-profiles/) as an archival feature.
  • [SAFE]: The skill's primary functionality is text processing and stylistic adaptation, which is performed through Claude's reasoning capabilities without external code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 12:34 AM