The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The script scripts/stack-health.sh programmatically extracts a Telegram bot token from the local configuration file ~/docker/stacks/monitoring/alertmanager/config.yml for use in API calls.
[REMOTE_CODE_EXECUTION]: Multiple commands in SKILL.md, references/alert-rules-guide.md, and scripts/stack-health.sh pipe output from curl directly to python3 -c for parsing. While the Python scripts are provided as static strings in the command line, this pattern involves executing code against data retrieved from network endpoints.
[COMMAND_EXECUTION]: The skill provides instructions and scripts that execute shell commands, manage Docker containers via docker compose, and interact with service APIs using curl.
[INDIRECT_PROMPT_INJECTION]: The skill processes and displays data from Prometheus and Alertmanager APIs. If an attacker can control metric labels or alert names in the monitoring system, they could potentially inject malicious instructions into the agent's context.
Ingestion points: Data is ingested from local Prometheus endpoints (/api/v1/rules, /api/v1/alerts) and the Telegram bot API.
Boundary markers: None identified; output from these services is processed and printed directly.
Capability inventory: The skill has the ability to execute shell commands, manage system services via Docker, read local configuration files, and perform network operations.
Sanitization: None; the scripts parse raw JSON data and print specific fields to the terminal without escaping or validation.

monitoring-stack