skills/zhiyuzi/hivo/hivo-drop/Gen Agent Trust Hub

hivo-drop

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates file management by executing local CLI commands through the hivo utility, including hivo drop upload, hivo drop download, hivo drop delete, hivo drop list, and hivo drop share.
  • [DATA_EXFILTRATION]: The skill's primary function is to upload local files to the external Hivo Drop service (hivo.ink). It also requires checking for the presence of a local authentication file, .hivo/identity.json, which is managed by the vendor's CLI.
  • [EXTERNAL_DOWNLOADS]: Fetches files from the remote Hivo Drop storage to the local machine or terminal output via the hivo drop download command.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it processes and can display the contents of remote files that could contain malicious instructions.
  • Ingestion points: Remote file content is read into the agent's context using the hivo drop download <path> command (when local save path is omitted).
  • Boundary markers: None identified; the skill does not specify any delimiters or warnings to ignore instructions found within downloaded file data.
  • Capability inventory: The skill has the capability to execute shell commands (hivo CLI), write to the local filesystem, and read local files.
  • Sanitization: No validation or sanitization of downloaded content is performed before it is processed or output to the terminal.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 04:36 AM