The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documents the usage of gfql_remote and python_remote_table which are intended library functions for executing graph queries and Python lambdas on a remote Graphistry server. As these are documented SDK features from a recognized vendor, they represent normal functionality for the skill's domain.
[PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface where user-derived query fragments are processed by the agent.
Ingestion points: User-provided strings used in g.gfql() and rg.gfql_remote() as documented in SKILL.md.
Boundary markers: The provided code snippets do not include specific delimiters, but the accompanying text instructs users to validate inputs.
Capability inventory: The skill allows for remote query execution and remote Python code execution (lambdas).
Sanitization: Recommends the use of parse_cypher and compile_cypher from the Graphistry library to validate and normalize input before execution.
[SAFE]: No obfuscation techniques, credential exposure, or unauthorized persistence mechanisms were detected. The skill's behavior is consistent with its stated purpose of assisting with Graphistry GFQL query construction.

pygraphistry-gfql