sift-codebase-audit

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust 'Non-negotiable operating contract' that strictly enforces read-only behavior, forbidding any edits to the repository, execution of tests, or network operations.
  • [COMMAND_EXECUTION]: The skill uses standard local inspection tools such as git, cat, and grep to analyze the codebase, which is consistent with its intended purpose and restricted to read-only operations.
  • [EXTERNAL_DOWNLOADS]: Installation instructions utilize npx and git clone to fetch the skill from the vendor's official GitHub repository (2389-research/sift), representing standard installation procedures from a known source.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 05:38 PM
Security Audit — agent-trust-hub — sift-codebase-audit