dingtalk-event

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the dws command-line utility to interact with DingTalk services. It executes commands such as dws event consume for streaming events, dws event schema for data structure inspection, and dws chat message send for responding to messages.\n- [PROMPT_INJECTION]: The skill processes real-time message content from DingTalk, creating an indirect prompt injection surface where external users could potentially send messages containing instructions for the agent.\n
  • Ingestion points: External message content is ingested via the dws event consume command in SKILL.md and references/event-im.md.\n
  • Boundary markers: The instructions rely on NDJSON formatting but do not include explicit delimiters or instructions for the agent to ignore commands embedded within the message content.\n
  • Capability inventory: The agent has access to several tools including dws chat message send, dws chat search, and dws aisearch person.\n
  • Sanitization: No sanitization or filtering of the message content field is mentioned in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 03:35 AM
Security Audit — agent-trust-hub — dingtalk-event