The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of external tools and packages.
Installs the @anthropic-ai/claude-code CLI, which is an official tool from Anthropic used as the coding agent engine.
Installs the agentic-data-scientist tool using the uv package manager.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data files which could potentially contain malicious instructions intended to influence the agentic workflow.
Ingestion points: Data is ingested via the --files and -f arguments, supporting CSV, JSON, and entire directories (SKILL.md).
Boundary markers: The instructions do not explicitly detail specific prompt delimiters or "ignore instructions" wrappers for the ingested data.
Capability inventory: The framework includes a 'Coding Agent' capable of implementing stages, which involves executing code and potentially shell commands via Claude Code (SKILL.md).
Sanitization: The skill uses a multi-agent review process (Plan Reviewer, Review Agent) to validate work, providing a layer of architectural oversight, though specific data sanitization routines are not detailed in the user-facing documentation.
[DYNAMIC_EXECUTION]: The core functionality relies on a 'Coding Agent' that dynamically generates and executes code to perform data science tasks.
The coding agent has access to 380+ scientific computing skills and executes implementation stages based on a plan derived from the user's query and the data provided.

agentic-data-scientist