The Agent Skills Directory

[DYNAMIC_EXECUTION]: The documentation in references/io.md includes examples of using pickle.load() and nx.read_gpickle() for graph serialization. Python's pickle module is fundamentally insecure as it can execute arbitrary code during the deserialization of data from untrusted sources.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from various external sources such as CSV, JSON, and GraphML files, which creates a potential surface for indirect prompt injection. Ingestion points: Multiple functions in references/io.md (e.g., read_edgelist, read_graphml, read_json) load external data into the agent's context. Boundary markers: No explicit instructions or delimiters are provided to the agent to distinguish untrusted data from instructions. Capability inventory: The skill permits file reading/writing and complex data processing. Sanitization: No data validation or sanitization steps are documented for the external data before it is processed.
[EXTERNAL_DOWNLOADS]: The skill documentation suggests installing well-known and trusted Python libraries such as networkx, pandas, and matplotlib from official package registries.

networkx