The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of external JSON data which may contain untrusted strings from user-supplied configurations.
- Ingestion points: The scripts slo_designer.py, alert_optimizer.py, and dashboard_generator.py ingest user-provided JSON files through the --input command-line argument.
- Boundary markers: The scripts do not implement specific delimiters or instructions to ignore embedded commands within the processed JSON fields.
- Capability inventory: The toolset includes file-writing capabilities via the --output and --report arguments across all included scripts.
- Sanitization: No explicit sanitization or escaping of the input data is performed, relying solely on standard JSON parsing.

observability-designer