The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to write and execute Python scripts to perform data analysis, enrichment, and variant interpretation based on retrieved biological data.
Evidence: SKILL.md section "COMPUTE, DON'T DESCRIBE" and DIAGNOSTIC_WORKFLOW.md providing code templates for execution.
[EXTERNAL_DOWNLOADS]: The skill fetches the esm library from a specific version on EvolutionaryScale's GitHub repository to enable variant mechanism analysis.
Evidence: DIAGNOSTIC_WORKFLOW.md and TOOLS_REFERENCE.md contain instructions to run pip install 'esm @ git+https://github.com/evolutionaryscale/esm@ee891c52'.
[COMMAND_EXECUTION]: The skill uses shell commands via Bash to run generated Python scripts and interact with the local file system for report generation.
Evidence: SKILL.md instructions to "write and run Python code via Bash".
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from various external sources, including patient-provided symptoms and automated searches of public databases like PubMed and BioRxiv, which could contain malicious instructions.
Ingestion points: Patient symptoms (SKILL.md), search results from Orphanet, OMIM, PubMed, and preprint servers (TOOLS_REFERENCE.md).
Boundary markers: Absent; there are no specific delimiters or "ignore" instructions for external data processing.
Capability inventory: The agent can execute shell commands and Python code, and write files to the local system (SKILL.md).
Sanitization: Absent; no sanitization or validation of data retrieved from external sources is described before it is processed or included in reports.

tooluniverse-rare-disease-diagnosis