nuke-cursor-app

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands like ps, pkill, osascript, and open to manage the lifecycle of the Cursor IDE application. These commands are restricted to the application's specific bundle path (/Applications/Cursor.app).
  • [INDIRECT_PROMPT_INJECTION]: The skill reads from a local metrics database to provide a summary of resource usage before and after the restart.
  • Ingestion points: Data is read from ~/Library/Application Support/<metrics-collector>/cursor-metrics.sqlite3 via sqlite3 commands in SKILL.md.
  • Boundary markers: None identified.
  • Capability inventory: The skill has the ability to terminate processes (pkill), execute AppleScript (osascript), write to the file system (mkdir), and execute arbitrary shell commands.
  • Sanitization: The skill uses specific SQL column selection (timestamp, role, process_count, cpu_percent, resident_bytes) to limit the scope of ingested data to numeric and status fields, reducing the likelihood of malicious text being processed as instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 01:44 PM
Security Audit — agent-trust-hub — nuke-cursor-app