knowledge-graph
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from a database, presenting an indirect prompt injection surface.
- Ingestion points: In Step 3, the skill queries the
kb_knowledgetable and extracts the unstructuredtextfield to parse for article numbers and relationships. - Boundary markers: The prompt does not define structural delimiters or specific instructions for the agent to isolate or ignore execution commands that might be embedded within the retrieved article text.
- Capability inventory: The skill possesses modification capabilities, including the
SN-Create-RecordandSN-Update-Recordtools, which could be abused if an injected instruction influences the agent's decisions. - Sanitization: There are no explicit sanitization or validation mechanisms outlined to filter out malicious patterns from the database records before processing.
Audit Metadata