lark-workflow-standup-report
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
lark-clibinary to authenticate and retrieve calendar events (+agenda) and task lists (+get-my-tasks). - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes user-generated content from an external platform (Lark/Feishu).
- Ingestion points: Untrusted data enters the context through the output of
lark-cli calendar +agenda(event summaries) andlark-cli task +get-my-tasks(task titles) inSKILL.md. - Boundary markers: The prompt template for the summary does not use explicit delimiters (e.g., XML tags or triple quotes) to separate fetched data from instructions, nor does it explicitly tell the agent to ignore instructions embedded in the data.
- Capability inventory: The skill possesses the capability to execute shell commands via
lark-cliacross its workflow. - Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the AI for the summary.
Audit Metadata