things3

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes shell commands via python3 and osascript to query the Things 3 database and trigger actions like adding or trashing tasks through the application's URL scheme and AppleScript interface.
  • [DATA_EXFILTRATION]: Accesses the local Things 3 SQLite database located at ~/Library/Group Containers/JLMPQHK86H.com.culturedcode.ThingsMac/ThingsData-*/Things Database.thingsdatabase/main.sqlite to read task and project information. This behavior is necessary for the skill's core purpose of task management.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it processes task titles and notes retrieved from the local database, which could contain content from external or shared sources.
  • Ingestion points: Task titles and notes are read directly from the TMTask table in the SQLite database.
  • Boundary markers: There are no specific delimiters or instructions provided to the agent to prevent it from obeying instructions embedded in the retrieved task content.
  • Capability inventory: The skill possesses the capability to execute Python scripts and AppleScript commands on the local host.
  • Sanitization: The skill does not perform sanitization or filtering on the retrieved database content before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 01:57 AM
Security Audit — agent-trust-hub — things3