The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides extensive documentation and tooling for interpolating external data into prompts, which inherently creates an attack surface for indirect prompt injection.\n
Ingestion points: Data enters the system through test_case.input in the scripts/optimize-prompt.py utility and via various template placeholders defined in references/prompt-templates.md and assets/prompt-template-library.md.\n
Boundary markers: The skill advocates for the use of clear step markers, numbered instructions, and explicit delimiters to maintain output consistency (see references/chain-of-thought.md).\n
Capability inventory: The scripts/optimize-prompt.py utility allows for automated interaction with LLM providers to evaluate prompt quality and efficiency.\n
Sanitization: The references/prompt-templates.md file provides a ValidatedTemplate implementation that demonstrates how to validate input variables (type, range, and enum choices) before they are rendered into the final prompt context.

prompt-pro