adk
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides installation commands for the Botpress ADK CLI that fetch and execute shell scripts from Botpress's official GitHub repository.
- Evidence:
curl -fsSL https://github.com/botpress/adk/releases/latest/download/install.sh | bashinSKILL.md. - Evidence:
powershell -c "irm https://github.com/botpress/adk/releases/latest/download/install.ps1 | iex"inSKILL.md. - [PROMPT_INJECTION]: The skill documents patterns for creating bots that process external data into knowledge bases for AI agents, creating an attack surface for indirect prompt injection.
- Ingestion points:
src/knowledge/usingDataSource.Websiteto index external URLs inSKILL.md. - Boundary markers: No explicit boundary markers or instruction isolation shown in code examples.
- Capability inventory: Supports sending messages and executing integration actions (Slack, GitHub, Linear) in
SKILL.md. - Sanitization: No input validation or sanitization shown for external content in
SKILL.md.
Audit Metadata