The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides command templates using curl to interact with an internal management API. These commands are used for legitimate administrative tasks like syncing monitors and triggering re-fetches. The use of environment variables such as $RELEASES_API_KEY for authentication is consistent with secure credential management practices.
[SAFE]: The skill contains no evidence of prompt injection, data exfiltration, or obfuscation. It focuses on internal operations for an 'Operator skill' and explicitly warns against misusing the production bindings in staging environments to prevent unintended data mutation.
[SAFE]: Instructions for parsing JSON using Python's json.load(..., strict=False) are included to handle non-standard characters in monitoring logs, which is a common technical requirement for handling raw web data and does not represent a security risk in this context.

firecrawl-monitoring