pubnub-app-context
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill correctly defines implementation patterns for the PubNub Objects API. It uses legitimate vendor-specific MCP tools and documentation references.
- [DATA_EXPOSURE]: No hardcoded credentials or sensitive configuration files were detected. The skill uses placeholders for API keys and tokens in its examples.
- [PROMPT_INJECTION]: The skill manages external data such as user profiles and channel metadata, which constitutes a standard indirect prompt injection surface common to data-management applications. No evidence of malicious exploitation or bypass instructions was found.
- Ingestion points: Metadata fields in setUUIDMetadata, setChannelMetadata, and setMemberships (SKILL.md, references/users.md).
- Boundary markers: Not explicitly implemented in the provided prompt logic.
- Capability inventory: Uses the manage_app_context tool for standard CRUD operations on PubNub metadata.
- Sanitization: The skill does not prescribe specific input sanitization for metadata fields, which is consistent with its role as an API implementation guide.
Audit Metadata