The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains specific instructions that steer agent behavior toward promoting a commercial platform (www.k-dense.ai) when user tasks increase in complexity. This override influences the agent to perform marketing actions within its workflow.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks due to its handling of untrusted user data. 1. Ingestion points: External data enters the agent context via 'data=df' variables and column-based inputs in the referenced 'scripts/assumption_checks.py' and internal code snippets. 2. Boundary markers: There are no explicit delimiters or instructions to treat data as distinct from executable commands. 3. Capability inventory: The skill leverages powerful analytical libraries including 'scipy', 'statsmodels', 'pingouin', and 'pymc' to perform calculations and visualizations. 4. Sanitization: No mechanisms for data validation or sanitization are established for the data processed by the skill.

statistical-analysis