The Agent Skills Directory

[SAFE]: The skill is entirely educational and provides secure coding patterns for logging and monitoring. It includes explicit warnings against logging sensitive information like passwords and API keys.
[CREDENTIALS_UNSAFE]: No hardcoded credentials were detected. The skill correctly demonstrates the use of environment variables (e.g., process.env.PAGERDUTY_KEY) for managing sensitive integration keys.
[COMMAND_EXECUTION]: The skill does not contain any shell command execution, privilege escalation attempts, or dangerous system-level operations.
[EXTERNAL_DOWNLOADS]: The code snippets reference standard, well-known Node.js libraries such as winston for logging and redis for state management, which are appropriate for the stated purpose.
[PROMPT_INJECTION]: There are no detected instructions intended to bypass safety filters, override system prompts, or extract internal agent configuration.
[DATA_EXFILTRATION]: While the skill demonstrates sending logs to external services (Splunk, Slack, PagerDuty), these are standard integrations for monitoring systems and are presented as examples using placeholders (e.g., splunk.example.com).

logging-monitoring