The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation includes a source link to the author's GitHub repository for additional research skills. This reference points to the vendor's own project infrastructure and is provided for informational and documentation purposes. Evidence: https://github.com/aipoch/medical-research-skills in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: As a data analysis tool, the skill naturally ingests user-supplied datasets, representing a potential vulnerability surface. However, the skill explicitly incorporates an 'Instruction Pollution Defense' (as noted in the changelog) to mitigate risks from malicious instructions embedded within processed data. Ingestion points: Data frames and arrays passed to analysis functions in SKILL.md. Boundary markers: Documented defense against instruction pollution. Capability inventory: Local file system writes to save analysis results. Sanitization: Structural and type validation is performed by the underlying statistical libraries.
[COMMAND_EXECUTION]: The skill uses Python to write statistical outputs to the local filesystem in CSV format, which is the expected and documented behavior for a data processing utility. Evidence: results_df.to_csv() and significant_genes.to_csv() calls in SKILL.md.

data-stats-analysis