The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs operations by invoking the wecom-cli binary. Commands are constructed by passing JSON data as a single-quoted string argument. This pattern requires the agent to be cautious, as improperly escaped single quotes within user-provided data could lead to command injection in the host shell environment.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests data from external WeCom sheets which may contain adversarial instructions.
Ingestion points: The get_doc_content and sheet_get_info tools retrieve content and metadata directly from online documents and provide them to the agent.
Boundary markers: There are no documented boundary markers or instructions advising the agent to treat the retrieved Markdown content as data rather than instructions.
Capability inventory: The agent has the ability to write data, delete sub-sheets, lookup user IDs via wecomcli-contact, and send messages via wecomcli-msg, creating a significant surface for automated actions triggered by injected text.
Sanitization: The skill does not implement or require any sanitization, filtering, or validation of the content pulled from the WeCom sheets before the agent processes it.

wecomcli-sheet