The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python code in Step 7 that programmatically writes Markdown files to the local file system using the open() function. This capability allows the creation of files based on user-provided titles.\n- [DATA_EXFILTRATION]: Example templates in the methodology and data documentation sections include placeholders for database connection strings (e.g., postgres://prod-db/analytics.user_events). While these are examples, they highlight a pattern of data exposure if users include real credentials in the generated documentation.\n- [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection. (1) Ingestion points: Untrusted analytical results and business context are ingested during the documentation process in SKILL.md. (2) Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used in the documentation templates. (3) Capability inventory: The skill possesses file-writing capabilities in Step 7. (4) Sanitization: No sanitization or escaping of the user-provided content is performed before interpolation into the final document.

analysis-documentation