apple-calendar

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes osascript (AppleScript) to interact with the macOS Calendar application. User-provided inputs (e.g., event summaries, descriptions, dates) are passed as positional arguments to the scripts and accessed via the argv property in AppleScript. The use of quoted heredocs (<<'EOF') in the shell scripts prevents shell interpolation, effectively mitigating shell injection risks.- [DATA_EXPOSURE]: The skill accesses and displays local calendar data, which includes personal and potentially sensitive information such as meeting summaries and descriptions. This behavior is consistent with the skill's primary function and is performed locally without network transmission.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes data from calendar events (summaries, descriptions, locations) that could be controlled by external parties (e.g., through meeting invitations).
  • Ingestion points: scripts/cal-events.sh, scripts/cal-read.sh, and scripts/cal-search.sh.
  • Boundary markers: None; external data is presented as plain text within the agent's context.
  • Capability inventory: Subprocess execution for calendar modifications via scripts/cal-create.sh, scripts/cal-update.sh, and scripts/cal-delete.sh.
  • Sanitization: No sanitization or validation of the event content is performed prior to being returned to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:44 PM
Security Audit — agent-trust-hub — apple-calendar