The Agent Skills Directory

[SAFE]: The skill provides legitimate templates and instructions for production observability. No malicious patterns, persistence mechanisms, or unauthorized privilege escalations were detected. The implementations for logging, metrics, and tracing follow industry standard best practices.
[INDIRECT_PROMPT_INJECTION]: The skill implements middleware that captures external request metadata, such as HTTP headers and URL paths, to provide context for logs and metrics. This constitutes a standard monitoring attack surface and is implemented with appropriate safeguards such as path normalization. * Ingestion points: LoggingContextMiddleware in logging-config-template.py (request headers) and PrometheusMiddleware in metrics-config-template.py (URL paths). * Boundary markers: None present in the provided templates. * Capability inventory: The skill uses scoped Bash(python:*), Bash(docker:*), Edit, and Write tools for configuration tasks. * Sanitization: PrometheusMiddleware._normalize_path performs sanitization by replacing numeric and UUID segments with placeholders to reduce cardinality and mitigate label-based risks.

monitoring-setup