spark-recipe-draft-batch
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from email threads and drafts, which represents a potential attack surface.
- Ingestion points: Email thread content and draft metadata are read into the agent context via
spark threadandspark emails Drafts(SKILL.md). - Boundary markers: The skill does not define specific delimiters to separate email content from agent instructions.
- Capability inventory: The skill can modify drafts (
spark draft --edit) and delete emails (spark action moveToTrash). - Sanitization: No sanitization of email content is performed before processing.
- Assessment: This vulnerability surface is inherent to the skill's primary purpose of email management and does not indicate malicious intent.
Audit Metadata