engineering-feishu-integration-developer

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions and code examples define a logic flow that ingests untrusted data from the Feishu (Lark) platform.
  • Ingestion points: The skill processes data from event subscriptions, specifically the im.message.receive_v1 event which contains user-provided message content.
  • Boundary markers: The provided code snippets lack explicit delimiters or instructions to the agent to treat the embedded content as untrusted or to ignore any commands within it.
  • Capability inventory: The skill provides logic for sending messages, creating/managing approvals, and performing Bitable CRUD operations.
  • Sanitization: No explicit sanitization or filtering of the message.content is demonstrated before it is used to trigger downstream actions such as handleBotCommand.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 06:15 AM
Security Audit — agent-trust-hub — engineering-feishu-integration-developer