The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill processes user requirements for 'Report purpose and data requirements' to generate executable report logic. This creates a surface where a malicious user could attempt to manipulate the agent into generating code that accesses unauthorized data or performs unintended actions.
Ingestion points: SKILL.md defines report purpose and data requirements as primary inputs.
Boundary markers: No specific delimiters or 'ignore' instructions are provided to the agent for processing these inputs.
Capability inventory: The skill leverages frappe.db.sql for database access and provides templates for Script Reports which execute Python and JavaScript.
Sanitization: The skill includes a dedicated 'Guardrails' section and a 'Common Mistakes' table that provide defensive instructions, such as using frappe.db.escape() and checking user permissions with frappe.has_permission().
[Dynamic Execution] (LOW): The skill involves the generation of Python and JavaScript scripts from templates to implement report logic, which is a standard and documented feature of the Frappe framework.

frappe-reports