networkx-graph-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides numerous methods to ingest untrusted data from external files and databases, creating a surface for indirect prompt injection.
- Ingestion points: Methods like
read_edgelist,read_graphml,read_adjlist,read_gml,read_gexf,read_pajek,read_leda(Module 6 in SKILL.md and Part 1 in references/io_visualization.md) andpd.read_sql_query(references/io_visualization.md) are documented. - Boundary markers: The instructions do not provide explicit delimiters or warnings for the agent to ignore embedded instructions within processed graph data.
- Capability inventory: The skill documents multiple file-writing capabilities across
SKILL.md(Module 6, Module 7, Recipes) andreferences/io_visualization.md(e.g.,write_edgelist,savefig,to_sql,write_html). - Sanitization: There is no evidence of sanitization, validation, or escaping of external content before it is processed or interpolated into output files.
Audit Metadata