jlcpcb-ordering

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXPOSURE]: The skill accesses local configuration and database files at ~/.config/jlcpcb-order/private.json and ~/.config/jlcpcb-order/orders.sqlite3. These files are used to store and retrieve order-related data and private configuration settings.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to read and react to content from the JLCPCB website, creating a potential surface for indirect prompt injection if malicious content were embedded in the order pages.\n
  • Ingestion points: Web page DOM elements, specifically the order-check drawer and price panel on jlc.com and jlcpcb.com via Chrome DevTools Protocol (CDP).\n
  • Boundary markers: The instructions include specific Chinese text markers (e.g., 检测到您的订单还有, 去填写) as stop conditions, which serve as functional boundaries.\n
  • Capability inventory: The skill can execute multiple local scripts for submission and automation, such as submit_board_order.py and order_assistant.py.\n
  • Sanitization: There is no evidence of explicit sanitization or escaping of the website content before it is processed by the agent logic.\n- [COMMAND_EXECUTION]: The skill relies on executing various local utilities and scripts (e.g., launch_shared_chrome.sh, quick_order_china.sh) to perform its automation tasks. It also includes a script to install a local desktop assistant from a ZIP file provided in the local environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:24 PM