clade-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSMETADATA_POISONING
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides code snippets that import from the Node.js package
@claude-ai/sdk. This is not the official Anthropic SDK, which is published as@anthropic-ai/sdk. The use of an unverified package name in deployment templates could lead to the installation and execution of unofficial or potentially malicious code if a package with that name is present in public registries. - [METADATA_POISONING]: The skill exhibits a consistent naming inconsistency, using the term 'Clade' in its primary name (
clade-deploy-integration) and when referencing related skills (clade-install-auth,clade-observability,clade-prod-checklist), while using 'Claude' in the description, tags, and instructions. This discrepancy may be an attempt to evade keyword-based security monitoring or to impersonate official tools while maintaining a distinct, potentially deceptive identifier.
Audit Metadata