lark-base

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external multi-dimensional tables in Feishu/Lark Base, creating a surface for indirect prompt injection attacks.
  • Ingestion points: Data enters the agent's context through several record-reading commands including +record-list, +record-get, and +record-search, as well as dashboard data retrieval via +dashboard-block-get-data (detailed in references/lark-base-record-query-and-analysis-sop.md).
  • Boundary markers: The instructions do not define clear delimiters or boundary markers to separate retrieved record data from the agent's system instructions.
  • Capability inventory: The skill provides the agent with high-privilege capabilities, such as creating/deleting records, managing advanced permissions and roles (+role-update), and configuring automated workflows that can execute HTTPClientAction webhooks.
  • Sanitization: There is no explicit logic described for sanitizing record values or verifying the integrity of data processed from the Base environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 06:56 AM
Security Audit — agent-trust-hub — lark-base