inbox-triage
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection (Category 8). It instructs the agent to process unread emails provided by the user, which constitutes untrusted data ingestion.
- Ingestion points: The user is prompted to paste unread emails into the context via the block
[paste your emails here]. (File:SKILL.md) - Boundary markers: There are no boundary markers (like XML tags or clear delimiters) or instructions to ignore malicious directives embedded within the email text.
- Capability inventory: The skill generates text summaries and drafts. While it does not currently invoke dangerous tools or perform network operations, an attacker could attempt to influence the agent's output (e.g., creating a malicious draft or misclassifying urgent items).
- Sanitization: No sanitization or validation of the input text is performed before it is processed by the AI model.
Audit Metadata