The Agent Skills Directory

[SAFE]: The skill implements a robust and secure workflow management lifecycle. It includes a strict "Authoring Pipeline" in SKILL.md that enforces validation via the control.workflows.validate tool, schema-locking from registry details, and a "Reference Safety Gate" to prevent common configuration errors and injection vulnerabilities.\n- [SAFE]: Data handling is secured through the mandatory use of workspace secrets and explicit expression wrappers (={{ ... }}), preventing the hardcoding of credentials or accidental execution of raw directives. The skill explicitly forbids plaintext credentials in workflow payloads in SECURITY.md.\n- [SAFE]: All network operations and resource references (e.g., https://nocode.ascn.ai/mcp, https://ascn.ai/no-code/mcp-list) are directed to official domains belonging to the vendor (ascnai), posing no risk of unauthorized data exfiltration or interaction with untrusted entities.\n- [SAFE]: The skill uses a comprehensive error taxonomy (contracts/error-taxonomy.yaml) and clearly defined recovery playbooks to handle dependency failures and capability gaps safely, ensuring the agent provides actionable instructions to the user rather than attempting unsafe improvisations.

ascn-operator