linkfox-fastmoss-product-rank-top-selling

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains instructions that constrain agent behavior, specifically mandating that the agent 'cannot give up' on downloading an onboarding zip file if user authorization is required. This overrides the agent's ability to autonomously decide to stop a process.
  • [EXTERNAL_DOWNLOADS]: To resolve authentication issues, the skill directs the agent to download a zip file from https://agent-files.linkfox.com/skills/linkfox-onboarding/release.zip. This involves downloading external content from the vendor's domain for local execution.
  • [COMMAND_EXECUTION]: The skill uses a Python script (scripts/fastmoss_product_rank_top_selling.py) to handle API communication and data storage. The script performs local file writes and directory creation within the project workspace.
  • [DATA_EXFILTRATION]: The skill transmits the LINKFOX_AGENT_API_KEY and session metadata (SESSION_ID, MODE_ID) to the vendor's API gateway at tool-gateway.linkfox.com to perform its primary function.
  • [PROMPT_INJECTION]: The skill ingests untrusted data (e.g., product titles and shop names) from a third-party API. It lacks defined boundary markers or sanitization procedures for this data, creating a potential surface for indirect prompt injection if the source data is malicious.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 08:08 AM