The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze output from the get_schedule_history tool to classify failures. This introduces a potential surface where malicious content embedded in job failure messages or history metadata could influence agent behavior.
Ingestion points: Output from the get_schedule_history tool referenced in the workflow and examples.
Boundary markers: Absent. No instructions are provided to the agent to treat history data as untrusted or to use specific delimiters.
Capability inventory: The skill utilizes the update_schedule tool to modify configuration and includes curl examples for network operations.
Sanitization: Absent. There are no instructions to sanitize or validate the content retrieved from job history before processing.
[COMMAND_EXECUTION]: The skill documentation includes a curl example for performing a PUT request to the Crontap REST API. While the command is intended for configuration, it demonstrates network communication and the use of environment variables ($CRONTAP_CLIENT_ID, $CRONTAP_API_KEY) for authentication. The targeted domain api.crontap.com is a recognized vendor resource for this skill.

crontap-retries-and-idempotency