skills/jeremylongshore/claude-code-plugins-plus-skills/cursor-extension-integration/Gen Agent Trust Hub
cursor-extension-integration
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires Bash(cmd:*) permissions to manage extensions. It explicitly includes instructions for 'cursor --install-extension', 'cursor --uninstall-extension', and a destructive 'rm -rf' command in the troubleshooting reference to delete extension data.
- [REMOTE_CODE_EXECUTION]: The skill facilitates the installation of external code by providing commands to download extensions from the VS Code Marketplace or install from local VSIX files. While the suggested extensions are from well-known sources, the mechanism allows for the execution of any extension-based code.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its processing of user-provided project requirements. A malicious user or project file could influence the agent to install a compromised extension. Ingestion points: User requests for project-specific extension setups (documented in SKILL.md and references/examples.md). Boundary markers: Absent. Capability inventory: File system deletion ('rm -rf') and extension installation ('cursor --install-extension') via bash (documented in references/installing-extensions.md and references/troubleshooting.md). Sanitization: None present.
Audit Metadata