The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill configures Prometheus to automatically discover targets using Kubernetes annotations and file-based service discovery, creating a surface for ingesting untrusted data.
Ingestion points: kubernetes_sd_configs (pod/service) and file_sd_configs in SKILL.md.
Boundary markers: None specified to differentiate between metric data and potential control instructions in the metadata.
Capability inventory: The skill uses helm, docker-compose, promtool, and curl for infrastructure management and verification.
Sanitization: Employs relabel_configs with regular expressions for mapping, but these do not provide instruction sanitization.
[COMMAND_EXECUTION]: Provides various CLI commands for system setup, configuration validation, and troubleshooting.
Evidence: Usage of helm install, docker compose, promtool check, and curl for API status checks in SKILL.md.
[EXTERNAL_DOWNLOADS]: Fetches Prometheus monitoring components from the official Prometheus Community repository.
Evidence: helm repo add prometheus-community https://prometheus-community.github.io/helm-charts in SKILL.md.

prometheus-configuration