The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches and installs the datamol library along with optional cloud backends s3fs and gcsfs from official package registries. These are standard tools for scientific data management.
[COMMAND_EXECUTION]: Employs the Bash tool to manage the Python environment and install required dependencies as part of the skill's setup process.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted molecular data from external file formats and remote URLs, creating a vulnerability surface for indirect instructions.
Ingestion points: External data is ingested through dm.read_sdf, dm.read_csv, dm.read_excel, and the universal dm.open_df function across references/io_module.md and references/core_workflows.md.
Boundary markers: No explicit natural language delimiters or "ignore embedded instructions" warnings are provided for the data being parsed into the agent's context.
Capability inventory: The skill has access to Bash, Read, Write, and Edit tools, allowing for substantial environmental interaction based on processed inputs.
Sanitization: The skill incorporates built-in chemical validation through dm.sanitize_mol and dm.standardize_mol to ensure structural integrity and filter malformed molecular data.

datamol