scylladb-docs
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by ingesting user questions and external documentation content from ScyllaDB's web properties. While this presents an indirect prompt injection surface, the skill lacks dangerous capabilities such as file system access or command execution.
- Ingestion points: User queries and external documentation content from
docs.scylladb.com(referenced inSKILL.mdandreferences/topic-map.md). - Boundary markers: The skill does not define specific delimiters to separate external documentation data from its internal instructions.
- Capability inventory: The skill is restricted to information retrieval; it contains no file-write, network exfiltration (beyond standard browser/tool browsing), or subprocess execution capabilities.
- Sanitization: No explicit sanitization or filtering of external documentation content is implemented.
Audit Metadata