The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from Lark OKR entities, including objectives, key results, progress records, and comments, which could contain malicious instructions designed to manipulate agent behavior. \n- Ingestion points: Untrusted data enters the agent context through several commands described in SKILL.md and references, such as +cycle-detail, +comment-list, +comment-detail, and +progress-list. \n- Boundary markers: The instruction set lacks explicit boundary markers or instructions to the agent to treat data retrieved from Lark as potentially untrusted or to ignore embedded instructions. \n- Capability inventory: The skill possesses significant write capabilities, including creating and modifying OKR content, comments, and progress records across a user's account via the lark-cli. \n- Sanitization: There is no evidence of content sanitization, filtering, or validation for the data retrieved from Lark APIs before it is used to inform further agent actions.

lark-okr