The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides explicit instructions and code patterns for the AI agent to execute system-level shell commands.
SKILL.md: Instructs the agent to use terminal commands like cat to generate temporary files and stata-mp -b to execute them in batch mode.
references/external-tools-integration.md: Documents the use of Stata's shell and ! operators to perform OS tasks such as ls, rm, and chmod.
[REMOTE_CODE_EXECUTION]: The skill enables the download and execution of external code from unverified sources and other programming environments.
packages/package-management.md: Describes using net install to download and run Stata code directly from various GitHub repositories (e.g., nppackages/binsreg, rdpackages/rdrobust).
references/external-tools-integration.md: Provides comprehensive guides for using python: and rcall: blocks, which allow the execution of arbitrary Python and R code within the Stata environment.
[EXTERNAL_DOWNLOADS]: The guide documents methods for fetching external data and resources.
references/data-import-export.md: Shows how to use import delimited and copy to download files from remote web URLs.
[PROMPT_INJECTION]: The skill documents a significant surface for Indirect Prompt Injection (Category 8).
Ingestion points: Data enters the agent's context from untrusted external CSV, Excel, and Web sources via the import and use commands documented in references/data-import-export.md.
Boundary markers: The reference files do not provide delimiters or instructions to ignore embedded commands in ingested data.
Capability inventory: The skill provides access to the shell, Python, and R environments, along with the ability to write files to disk (SKILL.md, references/external-tools-integration.md).
Sanitization: There is no mention of sanitizing or escaping ingested data before it is used in calculations or system calls.

stata