The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the networkx package and its optional dependencies using uv pip in SKILL.md. These are well-known, legitimate packages from the official Python Package Index.\n- [REMOTE_CODE_EXECUTION]: The skill documents the use of pickle.load() and nx.read_gpickle() for graph serialization in references/io.md. The pickle module is inherently insecure and can be exploited to execute arbitrary code if used to deserialize untrusted data provided by an external source.\n- [PROMPT_INJECTION]: The skill documents numerous functions for reading data from external sources such as GraphML, JSON, CSV, and SQL databases in references/io.md but does not provide guidance on input sanitization. This presents a surface for indirect prompt injection where malicious instructions could be embedded in node or edge attributes to influence agent behavior.\n
Ingestion points: nx.read_edgelist, nx.read_graphml, nx.read_gml, nx.read_json, and pd.read_sql_query in references/io.md.\n
Boundary markers: Absent.\n
Capability inventory: Includes file system write capabilities and visualization generation functionality.\n
Sanitization: Not documented.

networkx