Research MCP-Skillset Integration

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill consists entirely of instructional text and workflow diagrams. It does not include any scripts, executable commands, or remote code downloads.
  • [PROMPT_INJECTION]: The skill defines a process for ingesting external data from web and file sources, creating a standard surface for indirect prompt injection. Ingestion points: WebSearch, WebFetch, Grep, and Read (SKILL.md); Boundary markers: Not specified in instructions; Capability inventory: File system access, network access, and specialized code analysis; Sanitization: None specified. This is identified as a safe operational surface for this type of research skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 10:03 AM
Security Audit — agent-trust-hub — Research MCP-Skillset Integration