Skills
Navigate
skills/linkfox-ai/linkfox-skills/linkfox-temu-category-search/Gen Agent Trust Hub

linkfox-temu-category-search

Fail

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: HIGHCOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions in references/onboarding.md direct the agent to guide users in modifying shell configuration files (~/.zshrc, ~/.bashrc, ~/.bash_profile) and Windows environment variables via setx to persist API keys. These actions create persistence mechanisms for credentials across sessions. Additionally, scripts/temu_category_search.py dynamically modifies the Python module search path (sys.path.insert) to include relative external directories.
  • [CREDENTIALS_UNSAFE]: The script scripts/onboarding.py contains a hardcoded internal identifier (_LOGIN_FIXED_UID) used in authentication headers, which is a significant risk for credential management and potential spoofing.
  • [EXTERNAL_DOWNLOADS]: The onboarding script requires third-party Python packages requests, qrcode, and pillow which are not included in the standard library and lack version pinning, creating a dependency on external package registries.
  • [PROMPT_INJECTION]: The skill ingests user keywords in scripts/temu_category_search.py without sanitization, sending them to a remote API. The agent then processes the API response without boundary markers, creating an indirect prompt injection surface.
  • Ingestion points: keyword parameter in scripts/temu_category_search.py.
  • Boundary markers: None identified.
  • Capability inventory: Network requests via urllib.request.urlopen and local file writes to the project linkfox/ directory.
  • Sanitization: None.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 17, 2026, 07:03 AM
Security Audit — agent-trust-hub — linkfox-temu-category-search