cheat-recommend

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses read-only tools (Read, Glob, Grep) to analyze local markdown files (candidates.md, predictions/*.md) and state files. No write or execution capabilities are requested.
  • [SAFE]: No network operations, remote downloads, or external dependencies are present. All operations are confined to the local filesystem.
  • [SAFE]: Directory traversal patterns (../../shared-references/) are used to access configuration schemas and protocols within the expected project structure, presenting no security risk in this context.
  • [SAFE]: The workflow involves internal logic for filtering and sorting data without executing any dynamic or untrusted code.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 04:05 AM
Security Audit — agent-trust-hub — cheat-recommend