The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process chemical libraries (SMILES strings) and external CSV files, which are untrusted data sources. While these ingestion points could theoretically be used for indirect prompt injection attacks, the risk is negligible as the skill uses structured chemical parsers and lacks dangerous capabilities like network access, shell execution, or system modification.
Ingestion points: SMILES string lists in SKILL.md and examples/scaffold_split.py; CSV file references in usage-guide.md prompts.
Boundary markers: Absent in the example prompts, although the structured nature of SMILES strings provides inherent syntax validation.
Capability inventory: Structural chemical analysis (RDKit) and data manipulation (Pandas); no network operations, subprocess spawning, or persistence mechanisms are present.
Sanitization: The use of Chem.MolFromSmiles ensures that input strings conform to chemical grammar before analysis.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions to install standard, well-known scientific libraries including rdkit, mmpdb, scikit-learn, and datamol via standard package managers. These references are appropriate for the skill's stated purpose.

bio-scaffold-analysis