cursor

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill details the usage of the cursor CLI for various editor-related tasks such as opening files, using diff mode, and installing extensions.
  • [DATA_EXPOSURE]: Provides documentation on where Cursor stores user settings and keybindings on macOS and Linux (e.g., ~/Library/Application Support/Cursor/User/settings.json), as well as the location of the SSH configuration file (~/.ssh/config). This is standard documentation for the tool's environment and helps users navigate their local configuration.
  • [INDIRECT_PROMPT_INJECTION]: The skill documents a data ingestion surface where the agent can open local files or accept input from stdin (cursor -). 1. Ingestion points: File paths passed to the CLI and data piped via stdin (identified in SKILL.md and COMMANDS.md). 2. Boundary markers: Absent; the tool interprets inputs as file content or command arguments directly. 3. Capability inventory: Includes opening/diffing files and managing editor extensions (identified in SKILL.md and COMMANDS.md). 4. Sanitization: Absent; the CLI tool operates on provided inputs without intrinsic content filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 09:51 AM