The Agent Skills Directory

[SAFE]: The skill provides architectural guidance for project development using LLMs, advocating for staged pipelines and file-system state management.\n- [PROMPT_INJECTION]: The scripts/pipeline_template.py file defines a pipeline surface that processes external data from untrusted sources (APIs/Files). This implementation represents a surface for indirect prompt injection due to a lack of sanitization and strict boundary delimiters in the prompt templates.\n
Ingestion points: The fetch_items_from_source function in scripts/pipeline_template.py is designed to pull data from external sources.\n
Boundary markers: The PROMPT_TEMPLATE uses markdown headers but lacks strict delimiters (like XML tags or triple-backticks with 'ignore' instructions) for the interpolated {content} field.\n
Capability inventory: The script contains file system write operations and LLM invocation logic.\n
Sanitization: No sanitization or escaping of external content is present in the implementation template.

project-development