search
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFE
Full Analysis
- Search Orchestration: The skill provides a structured method for querying multiple connected services in parallel. This is a standard implementation for multi-source search tools.
- Data Retrieval Patterns: It translates user intent into platform-specific queries for chat, email, and storage tools. The use of MCP-connected sources ensures that data access is managed through established interfaces.
- Information Synthesis: Results are ranked and deduplicated before being presented as a unified answer. This approach focuses on providing relevant information to the user within the agent context.
- Indirect Prompt Injection Surface: Since the skill reads content from external sources such as emails and chat messages, there is an inherent surface for indirect prompt injection. This is a common property of tools that process external data. Evidence: (1) Ingestion points: External content from connected MCP tools like emails, chat logs, and documents (referenced in SKILL.md). (2) Boundary markers: Not explicitly specified in the prompt template. (3) Capability inventory: Search and read operations on various workspace tools. (4) Sanitization: The skill focuses on synthesizing and ranking results for final user presentation.
Audit Metadata