lsp-setup

Warn

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple shell commands to detect the operating system environment (uname -s), scan files via glob patterns, and verify or install software binaries.
  • [EXTERNAL_DOWNLOADS]: The skill initiates downloads of language server binaries and plugins from various sources including Homebrew (brew install), NPM (npm install -g), .NET Global Tools (dotnet tool install), and official GitHub releases. These are standard developer tools and official plugins for the agent environment.
  • [DATA_EXFILTRATION]: The skill reads global and project-specific configuration files (~/.claude/settings.json, .claude/settings.json) and system shell profiles (~/.bashrc, ~/.zshrc). While primarily used for state detection and enabling features, accessing these files presents a surface for potential data exposure if redirected.
  • [PRIVILEGE_ESCALATION]: Documentation in references/gotchas.md suggests that users configure passwordless sudo (NOPASSWD: ALL in /etc/sudoers) to avoid hang-ups during automated installation. This is a high-risk system modification that reduces the security posture of the host machine.
  • [PERSISTENCE_MECHANISMS]: The skill modifies shell profiles (~/.bashrc, ~/.zshrc) to add environment variables (ENABLE_LSP_TOOL=1) and update the system PATH. These changes persist across user sessions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 16, 2026, 09:36 PM
Security Audit — agent-trust-hub — lsp-setup