daily-brief
Pass
Audited by Gen Agent Trust Hub on Feb 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill handles untrusted data from multiple external sources which could contain malicious instructions meant to hijack the agent's behavior.
- Ingestion points: The variables
slack_messages,jira_updates,support_tickets, andcommunity_activityinSKILL.mdare directly interpolated into the prompt. - Boundary markers: There are no delimiters (like triple quotes or XML tags) or explicit instructions telling the model to ignore potential commands hidden within the provided context.
- Capability inventory: The skill currently only generates text, but the risk would escalate if the agent were granted additional tools (e.g., file-write or network-access).
- Sanitization: The input data is not sanitized or validated before being placed into the
Output Templatesection.
Audit Metadata