The Agent Skills Directory

[DATA_EXPOSURE]: The skill implements a security policy regarding sensitive data handling. It explicitly forbids the agent from using 'secrets' from the platform's Credential Safe as variables in pipeline relays, ensuring that credentials are not leaked into logs or transferred insecurely between execution steps.
[INDIRECT_PROMPT_INJECTION]: The skill handles data that could theoretically contain instructions (case descriptions and metadata). It identifies this attack surface and mitigates it through strict operational guidelines: requiring human verification for any orchestration logic inferred from external case metadata ('Fallback' mode) and providing clear boundaries between data fields and executable orchestration rules.

testany-pipeline