The Agent Skills Directory

Indirect Prompt Injection (SAFE): The script scripts/ascii_to_diagram.py parses markdown files to extract ASCII diagrams. This data ingestion is a known surface, but the script implements security controls including strict file extension validation and regex-based filename sanitization.
Ingestion points: Markdown files read from the local directory by scripts/ascii_to_diagram.py.
Boundary markers: Extracted ASCII content is clearly delimited within the generated report for the agent.
Capability inventory: The script is limited to file reading and report generation; it does not execute the extracted content.
Sanitization: Filenames are sanitized via regex, and input is limited to .md and .markdown files.
External Downloads (SAFE): The skill recommends installing standard, well-maintained packages (diagrams, matplotlib) and the Graphviz system utility. No suspicious third-party repositories or unverifiable binary downloads are referenced.
Command Execution (SAFE): The provided scripts use standard Python library calls for environmental checks and file operations. No dangerous subprocess calls or privilege escalation attempts were identified.

azure-diagrams