The Agent Skills Directory

[SAFE]: The skill implements a structured workflow with explicit boundaries and mandates human validation before creating or modifying high-level organizational structures. For example, it instructs the agent to 'propose name + telos and get the nod before creating' and to bring unresolved tensions to the user's agenda rather than resolving them autonomously. This process-level design reduces the risk of unauthorized state changes.
[COMMAND_EXECUTION]: The skill utilizes several specialized tools (e.g., nks_orient, nks_add_bianhua, nks_semantic_search) for interacting with a knowledge system. These operations are performed within the intended scope of the skill's functionality and do not involve arbitrary command execution or unauthorized shell access.
[DATA_EXPOSURE]: While the skill processes data from a 'realm' using tokens, these tokens are used as parameters for authorized tools. There is no evidence of data exfiltration to external domains or unauthorized credential harvesting.

assembly