The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx promptfoo@latest to download and run the Promptfoo evaluation tool. This is a standard and well-documented method for utilizing the framework.
[COMMAND_EXECUTION]: Provides instructions for running CLI commands such as eval, init, and view to manage the evaluation lifecycle.
[PROMPT_INJECTION]: Employs prompt templates with variable interpolation (e.g., {{user_input}}, {{task}}), which serves as a potential surface for indirect prompt injection if untrusted data is processed.
Ingestion points: Data is ingested from local files specified in tests/cases.yaml and prompt files like prompts/system.md.
Boundary markers: The skill uses standard double curly brace {{}} syntax for variable substitution but does not define explicit delimiters or escaping for interpolated content.
Capability inventory: The skill is capable of executing CLI commands and local Python scripts defined in the configuration (e.g., scripts/metrics.py).
Sanitization: No explicit sanitization or validation logic is provided for the input variables before they are interpolated into the prompts.

promptfoo-evaluation