The Agent Skills Directory

Data Ingestion Surface: The skill is designed to profile and summarize external datasets (CSV, JSON, Parquet) located in the workspace. While this is the primary purpose of the skill, it creates a surface for indirect prompt injection where untrusted data could potentially influence downstream processes if that data contains instructional content.
Ingestion points: The workflow in SKILL.md specifies scanning and loading all data files from the workspace directory.
Boundary markers: The current instructions do not specify explicit delimiters or "ignore instructions" warnings for the structured JSON output passed to other skills.
Capability inventory: The skill utilizes the pandas library for statistical profiling and workspace file access within the sandbox environment.
Sanitization: The skill primarily extracts metadata and statistical summaries (row counts, null counts, data types), which minimizes the risk of passing raw, potentially malicious strings to other components.

data-explorer