lark-calendar
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [METADATA_POISONING]: The skill YAML frontmatter lists the author as 'larksuite', which is inconsistent with the provided author context of 'seaworld008'. This mismatch could lead users to misinterpret the skill as an official Larksuite release.
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from both user inputs and API responses without explicit sanitization or boundary markers. * Ingestion points: Untrusted content is ingested through event summaries, descriptions, and search queries in SKILL.md and references/lark-calendar-create.md. * Boundary markers: The skill does not utilize delimiters or specific instructions to isolate these inputs from the system instructions. * Capability inventory: The skill has the capability to execute varied calendar operations through the 'lark-cli' binary and includes a feature to read local files if a path is prefixed with '@' in the description field. * Sanitization: The instructions lack requirements for validating or escaping the interpolated data strings before they are processed by the agent.
Audit Metadata