The Agent Skills Directory

[SAFE]: The skill provides standard architectural patterns and source code for alert management. All detected functionalities are consistent with the stated purpose of coordinating monitoring and incident response.
[CREDENTIALS_UNSAFE]: No hardcoded credentials or secrets were found. The implementation examples correctly use environment variable placeholders (e.g., PAGERDUTY_API_TOKEN, SLACK_WEBHOOK_URL, PAGERDUTY_SERVICE_KEY) to handle sensitive authentication tokens.
[EXTERNAL_DOWNLOADS]: The skill does not perform any unauthorized or suspicious external downloads. The provided docker-compose.yml references official images from the Prometheus organization (prom/prometheus, prom/alertmanager).
[DATA_EXFILTRATION]: Network operations are performed using the axios library and target established, well-known service domains for PagerDuty (api.pagerduty.com, events.pagerduty.com) and Slack webhooks. There is no evidence of sensitive local file access being combined with network transmission.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of alert summaries and custom details. While this represents a theoretical ingestion surface for untrusted data, it is a core functional requirement for the intended use case. The provided code does not expose the agent's internal control logic to this data.
[COMMAND_EXECUTION]: The inclusion of a validation script (validate-api.sh) is benign, as it currently contains only placeholder logic and does not execute dangerous shell commands or inject user input into sensitive sinks.

alert-management