whobot-phonecall-skill
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a proprietary command-line interface (
whobot) to perform agent management and telephony tasks. Execution is governed by a structured JSON protocol provided by the vendor's infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including audio recordings (
--recording), business documents (--document), and spreadsheets (--file). These ingestion points create a surface for indirect prompt injection where instructions embedded in the user-provided data could attempt to influence agent behavior. The skill mitigates this by instructing the agent to treat such content strictly as business data. - [EXTERNAL_DOWNLOADS]: The skill provides procedures for installing the vendor's official CLI tool via
npx --yes @whobot/cli@latest install, fetching the package from the standard public registry. - [DATA_EXFILTRATION]: The skill is designed to transfer local business assets and lead lists to the vendor's cloud platform for training and execution. The instructions require the agent to obtain explicit user confirmation before initiating these data transfers.
Audit Metadata