knowledge-graph-creation
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process arbitrary text provided by users or external documents to extract entities and relationships. This ingestion of untrusted data represents an attack surface where malicious instructions could be embedded within the source text to influence the agent's behavior. \n- Ingestion points: Workflow Step 1 and the 'Usage' section describe reading 'input text', 'document', or 'set of documents'. \n- Boundary markers: The instructions do not prescribe the use of delimiters (like triple quotes or XML tags) or specific system instructions to ignore embedded commands within the provided text. \n- Capability inventory: While the skill itself does not invoke tools, it generates structured code (Cypher, JSON-LD, Mermaid) which may be executed by the agent or downstream systems. \n- Sanitization: No sanitization or validation logic is defined to filter the content of the extracted entities or predicates.
Audit Metadata