morning-briefing
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including unread emails, Slack DMs, and news headlines to generate summaries and synthesis. It lacks explicit instructions for using boundary markers or sanitizing this data, creating a surface where malicious instructions hidden in an email or message could potentially influence the agent's behavior or output.
- Ingestion points: Data is pulled from Gmail, Slack, and news sources as described in the 'Gather the data' section of SKILL.md.
- Boundary markers: No delimiters or 'ignore' instructions are specified for external content.
- Capability inventory: The skill can write files to the local system (HTML dashboards and configuration files) and send data via Slack connectors.
- Sanitization: There is no mention of filtering or escaping external content.
- [PERSISTENCE]: The skill contains instructions to set up recurring scheduled tasks (cron-based) to automate the briefing process. While this is an intended feature of the 'Running on a schedule' section to facilitate daily reports, it involves creating persistent execution triggers on the system.
Audit Metadata