The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests extensive data from external biological sources, including MGI, ZFIN, FlyBase, WormBase, SGD, RGD, and GBIF. This data is used for multi-step reasoning and synthesis. There is a risk that malicious instructions could be embedded in the metadata or descriptions retrieved from these external registries. * Ingestion points: Results from tools such as MGI_get_gene, FlyBase_get_gene, WormBase_get_gene, ZFIN_get_gene, SGD_get_gene, and RGD_get_gene across all phases of the pipeline (Phase 1-5). * Boundary markers: Absent. The instructions do not specify the use of delimiters or specific warnings to ignore instructions within the retrieved biological data. * Capability inventory: The skill has the capability to generate and execute Python code and Bash commands. * Sanitization: Absent. There is no explicit requirement to sanitize or validate the content retrieved from external databases before it is processed or used in code generation.
[DYNAMIC_EXECUTION]: The 'COMPUTE, DON'T DESCRIBE' directive instructs the agent to write and run Python code via Bash to perform statistics, data processing, and scoring. This involves the dynamic generation of code based on the context of the analysis, which includes data from external tools. The libraries utilized include pandas, scipy, statsmodels, and matplotlib.

tooluniverse-model-organism-genetics