geo-agent

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection due to how it handles external data.
  • Ingestion points: scripts/competitor_research.py fetches raw data from Baidu and Bing search results, and scripts/index_checker.py ingests responses from AI platforms.
  • Boundary markers: Absent. The scripts/article_generator.py script interpolates raw competitor descriptions directly into the prompt without delimiters or instructions to ignore embedded commands.
  • Capability inventory: The skill has the capability to publish content to major social platforms (Zhihu, Baijiahao, Sohu, Toutiao) via scripts/publisher.py and interact with AI search engines.
  • Sanitization: No sanitization or filtering is performed on the scraped content before it is processed by the LLM.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts to automate its workflow.
  • It utilizes python3 to run management, research, and reporting modules as subprocesses.
  • It uses playwright to automate browser interactions for login and publishing.
  • [EXTERNAL_DOWNLOADS]: Fetches browser binaries from the official Playwright repository.
  • The installation process involves downloading necessary dependencies from Microsoft's Playwright infrastructure to enable browser automation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 07:42 PM