deadline-and-reply-tracking
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external information such as messages, calendars, and task records, creating an attack surface for indirect prompt injection. The instructions include robust mitigations for this risk.\n
- Ingestion points: Identified in
SKILL.mdas message bodies, invitations, and record notes.\n - Boundary markers: The skill requires wrapping retrieved material in
<EXTERNAL_DATA source="...">delimiters.\n - Capability inventory: The skill's focus is on reading and reconciling records; it explicitly denies external content the ability to authorize writes, command execution, or credential use.\n
- Sanitization: The instructions mandate ignoring embedded commands within untrusted data and strictly validating URLs to block access to private or local networks.\n- [NO_CODE]: The skill is composed of instruction files and configuration metadata without any accompanying executable scripts or binaries.
Audit Metadata