xiaoqi-fitness-tracker

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as food photos, fitness reports (InBody, Visbody), and application screenshots through OCR. While this presents a surface for indirect prompt injection if the ingested content contains adversarial instructions, the risk is inherent to the skill's primary function of data extraction.
  • Ingestion points: Processes user-provided images and reports in the '体测数据录入' and '饮食记录' functions.
  • Boundary markers: No specific delimiters or safety instructions are defined for the data extraction process.
  • Capability inventory: The skill possesses file read/write access to markdown files and HTML within the 05_健身/ directory.
  • Sanitization: The instructions do not specify validation or sanitization routines for extracted text.
  • [SAFE]: The skill maintains all operations within a dedicated local directory (05_健身/) and does not attempt to access sensitive system paths (e.g., SSH keys, environment variables) or external network resources.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 09:11 AM
Security Audit — agent-trust-hub — xiaoqi-fitness-tracker