dingtalk-workflow-social-tracker

Pass

Audited by Gen Agent Trust Hub on Apr 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the dws CLI tool to interact with DingTalk Aitable. This includes administrative operations like creating bases, tables, and fields, as well as data-level operations like creating, querying, and updating records. These commands are integral to the stated purpose of the skill.\n- [EXTERNAL_DOWNLOADS]: The skill incorporates an AI-driven WebSearch capability to retrieve public subscriber and engagement data from social media platforms and analytics sites like Social Blade. This data is used to populate daily snapshots.\n- [PROMPT_INJECTION]: The skill possesses an inherent attack surface for indirect prompt injection because it ingests and processes untrusted data from the open web to generate summaries and bot notifications.\n
  • Ingestion points: External data retrieved via WebSearch from social media platforms and third-party profile analytics pages.\n
  • Boundary markers: The skill does not implement specific boundary markers or 'ignore' instructions to isolate retrieved web data from its processing logic.\n
  • Capability inventory: The skill has the ability to write records to Aitable and send automated messages to users or groups via the send-by-bot command.\n
  • Sanitization: There is no documented logic for sanitizing or validating external content before it is parsed by the AI or written to the database.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 14, 2026, 02:53 AM