tencent-docs

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The OperationSheet tool in sheet/api/operation-api.md allows for the generation and execution of JavaScript-based scripts for complex spreadsheet manipulations. This is a core feature for advanced table editing and is managed through structured AI-generated commands.\n- [PROMPT_INJECTION]: The skill includes explicit defensive instructions in sheet/api/js-script-rule.md to prevent the AI model from responding to injection attempts or disclosing internal configurations. This acts as a security guardrail rather than a threat.\n- [EXTERNAL_DOWNLOADS]: The setup.sh script automates the installation of the mcporter tool via the global NPM registry. This is a legitimate administrative procedure required for the skill to communicate with the MCP host.\n- [INDIRECT_PROMPT_INJECTION]: The scrape_url and get_content features provide mechanisms for the agent to ingest external content from the web or other documents. This represents a standard document processing attack surface that is mitigated by the safety principles defined in the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 03:44 AM
Security Audit — agent-trust-hub — tencent-docs