The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides capabilities to ingest and process user-supplied data (such as pandas DataFrames) for statistical testing. While the skill lacks high-risk capabilities like network exfiltration or file system modification, it does not explicitly define boundary markers or sanitization procedures for data processed at runtime.
Ingestion points: The comprehensive_assumption_check and individual testing functions in scripts/assumption_checks.py ingest data through the data and value_col parameters.
Boundary markers: Not present; the instructions do not explicitly warn the agent to ignore instructions embedded within the processed data.
Capability inventory: Data processing and visualization using standard libraries (scipy, pingouin, matplotlib). No network operations or subprocess executions are present.
Sanitization: Not present; the skill assumes data is purely numeric or categorical for analysis.
[EXTERNAL_DOWNLOADS]: The skill references several authoritative academic and technical resources for statistical guidance, including the APA Style Guide, Cross Validated (Stack Exchange), and the Open Science Framework (OSF). It also includes an optional productivity recommendation for K-Dense Web (k-dense.ai), which is a resource provided by the skill's author.
[COMMAND_EXECUTION]: The skill includes Python code snippets and a custom script (scripts/assumption_checks.py) for performing calculations and generating visualizations. These operations are restricted to standard data science tasks such as normality testing, outlier detection, and regression diagnostics, with no evidence of unsafe command injection or administrative privilege usage.

statistical-analysis