lark-workflow-knowledge-base
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted external sources, including articles, YouTube video transcripts, Twitter threads, and PDF files, as well as Lark chat messages.
- Ingestion points: Content is retrieved from user-provided URLs in Step 2 and from Lark messages using
lark-cli im +chat-messages-listin Step 1. - Boundary markers: The instructions lack explicit boundary markers or "ignore embedded instructions" warnings for the AI when processing external content.
- Capability inventory: The skill possesses capabilities to create documents (
lark-cli docs +create), update database records (lark-cli base +record-upsert), and send messages to users (lark-cli im +messages-send). - Sanitization: There is no mention of sanitizing, escaping, or validating the external content before it is processed by the AI or used as arguments in shell commands.
Audit Metadata