The Agent Skills Directory

[SAFE]: The skill implements a rigorous triaging and verification process with multiple 'gates' (Evidence, Ship, and Decision) to ensure all system changes are tested and verified before deployment.- [SAFE]: Explicit guardrails are defined to prevent the accidental exposure of private data, credentials, and secrets in commands, logs, commits, and pull requests.- [SAFE]: The instructions mandate the use of synthetic examples and generic identifiers for any public artifacts, protecting the privacy of the underlying infrastructure.- [SAFE]: The skill promotes data integrity by utilizing a structured store with built-in validation commands (e.g., 'learnings validate') to detect inconsistencies or errors in the evidence log.

learn-loop