The Agent Skills Directory

[SAFE]: The skill provides utility functions for inventorying and classifying documents. The code uses standard libraries (pathlib, re, yaml) and safe methods for data handling.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest and process untrusted document content.
Ingestion points: The inventory_documents and extract_features functions in SKILL.md read the full text of files matching specific patterns from a provided root directory.
Boundary markers: The demonstration code does not implement specific boundary markers or "ignore instructions" delimiters for the content being read.
Capability inventory: The skill's capabilities are limited to file system traversal, file reading, regex extraction, and structured data parsing. It lacks network access, file-writing capabilities, or shell command execution.
Sanitization: The skill correctly uses yaml.safe_load for parsing frontmatter, which prevents unsafe object instantiation during metadata extraction.

document-audit-extraction