The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute local scripts such as scripts/validate_figure.py and scripts/plot_templates.py. These scripts are used for static preflight analysis of plotting code and for generating charts from data. The execution is scoped to the skill's directory and standard Python/R environments.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, specifically user-provided CSV files and manuscript legends. It mitigates potential injection risks by implementing strict data mapping contracts (asset-adaptation.md), numeric coercion in plotting scripts (plot_templates.py), and mandatory reporting of data exclusions and transformations.
[SAFE]: The skill includes references to official scientific publishing domains (e.g., nature.com, sciencedirect.com, rupress.org) for style and ethics guidance. It also maintains a provenance record for bundled image assets (assets/README.md), following established academic and open-source standards.

nature-figure