feishu-doc
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external Feishu documents (read, list_blocks) and provides the agent with powerful capabilities (write, append, delete_block, update_block) to modify or destroy data based on that content.
- Ingestion points: The skill ingests untrusted data through
action: "read"andaction: "list_blocks"inSKILL.md. - Boundary markers: No specific instructions or boundary markers are provided to the agent to ignore potentially malicious instructions embedded within the Feishu documents it reads.
- Capability inventory: The skill possesses extensive document manipulation capabilities, including
write(which replaces the entire document),append,update_block, anddelete_blockas defined inSKILL.md. - Sanitization: There is no evidence of content sanitization or validation before the agent acts on instructions derived from the document content.
Audit Metadata