wacli
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill metadata specifies the installation of the
waclibinary from a third-party repository on GitHub (github.com/steipete/wacli) using the Homebrew or Go package managers. - [COMMAND_EXECUTION]: The skill operates by executing various
waclisubcommands to perform actions such asauth,sync,chats list,messages search, andsend text/file. - [PROMPT_INJECTION]:
- Ingestion points: The skill ingests untrusted data from external WhatsApp messages via the
wacli messages searchandwacli chats listcommands. - Boundary markers: The instructions do not define specific delimiters or "ignore" directives to prevent the agent from being influenced by instructions contained within the retrieved WhatsApp messages.
- Capability inventory: The skill has the capability to send messages to arbitrary phone numbers or groups and can access local storage at
~/.wacli. - Sanitization: There is no explicit sanitization or escaping of the message content described in the skill's operational flow.
Audit Metadata