The Agent Skills Directory

[SAFE]: The skill defines a formal protocol for managing transformation wavefronts and auditing prose claims using domain-specific graph tools. The operations are confined to internal knowledge graph manipulation and do not involve system-level commands, network exfiltration, or sensitive file access.
[SAFE]: While the skill ingests external prose artifacts for auditing, the process is highly structured and read-only. The lack of dangerous primitives (such as shell access, file writes, or network requests) and the use of specific vendor-scoped tools (nks_*) minimizes the risk of indirect prompt injection.
[SAFE]: No obfuscation, hidden instructions, or unverifiable remote dependencies were detected in the skill instructions or metadata.

integrity