pp-braze

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install a CLI tool using standard package managers (NPM and Go) targeting the official repositories of the author, 'mvanhorn'.\n- [COMMAND_EXECUTION]: Executes the braze-pp-cli binary using the Bash tool to perform API operations and configuration tasks.\n- [DATA_EXFILTRATION]: The CLI utility includes a documented --deliver webhook:<url> feature that allows the output of commands (such as campaign data or analytics) to be sent to a user-specified URL. It also supports an optional remote feedback endpoint configured via environment variables.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external data from the Braze REST API (campaign metadata, names, and descriptions) and user-supplied strings via the which capability command. \n
  • Ingestion points: CLI command arguments and external Braze API results processed by the binary in the Bash environment.\n
  • Boundary markers: None provided in the skill instructions to separate data from commands.\n
  • Capability inventory: Full execution of the braze-pp-cli binary, which has network access and local file system write capabilities via the delivery sink.\n
  • Sanitization: No explicit sanitization or filtering is described in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 09:18 AM
Security Audit — agent-trust-hub — pp-braze