skills/lulu-eva/eva-skill/eva-review/Gen Agent Trust Hub

eva-review

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a local record-keeping system that stores content reviews in a structured directory (./eva-review/). This is standard functionality for maintaining a project-specific history.
  • [SAFE]: The skill includes specific privacy instructions in SKILL.md and references/review/02_single_单篇复盘.md to avoid copying raw private data from user materials (like screenshots or comments) into its summary cards, preferring minimal summaries instead.
  • [SAFE]: Data ingestion practices for external materials (CSV, Excel, and screenshots) are governed by a safety boundary reference and include instructions to only extract clearly visible facts without making unfounded guesses.
  • [SAFE]: The routing logic for different review modes (single, batch, backfill) is transparently documented in the instruction files and does not contain hidden or malicious commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 08:53 AM
Security Audit — agent-trust-hub — eva-review