contact-vcard-extractor

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security vulnerabilities were identified. The skill performs its stated function using local scripts and platform-provided utilities.
  • [COMMAND_EXECUTION]: The skill executes a bundled Python script (contact_to_vcard.py) to handle vCard generation and apple-vision for OCR. These are standard operations for the skill's purpose and occur within the local workspace.
  • [DATA_EXFILTRATION]: No network activity was detected. The skill reads from and writes to the local workspace and attachment directories without external communication.
  • [PROMPT_INJECTION]: The skill processes untrusted data from user messages and OCR results, which is an inherent surface for indirect prompt injection. However, the risk is mitigated by the use of a specialized parsing script that applies regex-based validation and proper vCard escaping.
  • Ingestion points: Reads text from user messages, workspace files, and OCR output from apple-vision.
  • Boundary markers: The skill does not use explicit delimiters for untrusted input, but instructions define strict field extraction and output formatting.
  • Capability inventory: Utilizes script execution, file writing, and image processing tools.
  • Sanitization: The Python script implements specific character escaping and regex filtering for contact fields (phone, email, URLs) to ensure vCard format integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 12:06 PM