skills/aws-samples/sample-agent-skills-for-builders/agentic-responsible-ai-assessment/Gen Agent Trust Hub
agentic-responsible-ai-assessment
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill allows users to provide their own questionnaire file path to replace the default assessment file. This introduces a surface for indirect prompt injection where instructions hidden in the user-supplied data could influence the agent's behavior. \n
- Ingestion points:
SKILL.md(via user-supplied questionnaire path).\n - Boundary markers: Absent. No instructions are provided to the agent to treat the file content strictly as data.\n
- Capability inventory: The file content drives the agent's logic, including the specific questions asked and the scoring criteria applied.\n
- Sanitization: Absent. No validation or filtering is performed on the external questionnaire content.\n- [DYNAMIC_CONTEXT_INJECTION]: The skill uses shell command execution to probe environment variables (e.g.,
$TERM,$VSCODE_PID,$SSH_TTY) to identify the host environment. This information is used to select a UI rendering mode (HTML, Mermaid, or ASCII). While intended for UI adaptation, this constitutes system fingerprinting and potential reconnaissance.\n- [DYNAMIC_EXECUTION]: The skill dynamically generates HTML and JavaScript artifacts at runtime to render Highcharts visualizations in supported clients. While it utilizes local vendor libraries, the runtime generation of executable UI content is a dynamic execution pattern.
Audit Metadata