wps

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of content from untrusted external documents and web sources, providing a surface for indirect prompt injection attacks.
  • Ingestion points: The skill reads document content via read_file_content (references/drive/read_and_download.md), scrapes web pages using scrape_url (references/drive/create_and_upload.md), and retrieves data from spreadsheets via sheet.get_range_data (references/sheet/data.md) and dbsheet.list_records (references/dbsheet/record.md).
  • Boundary markers: The instructions do not define delimiters or specific system prompts to isolate external content or warn the agent to ignore instructions found within that content.
  • Capability inventory: The skill provides significant capabilities including document modification, permission management, and file sharing through the kdocs-cli tool.
  • Sanitization: No evidence of content filtering or sanitization was found in the provided documentation.
  • [EXTERNAL_DOWNLOADS]: The skill downloads and installs an external binary tool as part of its setup process.
  • Evidence: Installation scripts in the scripts/ directory (setup.sh, setup.ps1, setup.cjs) download the kdocs-cli binary from https://wpsai.wpscdn.cn.
  • Context: The download source is an official CDN domain for Kingsoft WPS services, which is recognized as a well-known service provider.
  • [COMMAND_EXECUTION]: The skill relies on local shell command execution for its primary functionality and installation.
  • Evidence: The agent is instructed to use the kdocs-cli binary for all interactions with the document API. The setup scripts also utilize shell commands to extract archives, copy files, and modify the user's PATH variable.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 04:28 AM
Security Audit — agent-trust-hub — wps