linkfox-ai-mode-google-search
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external content from Google AI Search results which may contain untrusted data or hidden instructions.
- Ingestion points: Scraped AI Overview summaries provided in the
stdoutresponse field (SKILL.md,scripts/google_ai_search.py). - Boundary markers: The skill lacks explicit delimiters or instructions to the agent to disregard commands embedded within the search results.
- Capability inventory: The skill scripts perform network requests to vendor APIs and write session data to the project directory (
scripts/google_ai_search.py). - Sanitization: The skill processes raw Markdown text without specific sanitization against prompt injection.
- [PERSISTENCE]: The onboarding documentation (
references/onboarding.md) provides instructions for the agent to help users modify their shell profiles (~/.bashrc,~/.zshrc) to store theLINKFOX_AGENT_API_KEY. While intended for configuration, this involves modifying persistent system files. - [EXTERNAL_DOWNLOADS]: The skill's scripts reference external Python libraries (
requests,qrcode,pillow) and suggest their installation viapipif missing from the execution environment. - [COMMAND_EXECUTION]: The skill includes instructions for the user to run shell commands to set environment variables required for API authentication.
Audit Metadata