The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data sources, including CSV/TSV files, DataFrames, and rendered image panels (PNG/PDF), which introduces a potential surface for indirect prompt injection if the input files contain adversarial content.
Ingestion points: As described in SKILL.md, the agent is instructed to gather data from workspace-relative files or user-provided image panels to populate the figure.
Boundary markers: There are no explicit instructions to use delimiters or protective markers when interpolating data content into the agent's execution context.
Capability inventory: The skill generates scripts that perform file system writes (fig.savefig) and directory management (os.makedirs) within the figures/ workspace path.
Sanitization: The skill lacks specific guidelines for validating or sanitizing the content of the data files before processing them through the plotting functions.

multipanel