/rrr — GLM-5 Code Review

Trust Score 76/100

from dgk-claude16

Run a deep GLM-5 (744B) code review on staged, unstaged, or PR diffs with background execution and configurable modes.

triggers:/rrrglm-5code reviewdiffpr reviewstaged review

GitHub SKILL.md

What it does

/rrr runs an in-depth code review using a GLM-5 744B model. It prepares diffs (committed, staged, or PR), runs the review CLI in a background process, and returns structured findings for verification and remediation. The Skill documents safe execution patterns for multi-session environments.

When to use it

Use when you need a deeper, paid review than the lighter /rr flow — for thorough security-focused or PR-level reviews, or when operating in multi-session CI-like environments where diff-file input is required. Supports: staged reviews, PR diffs, and custom instruction prompts (e.g. 'security review').

What's included

Scripts: no repo scripts detected (has_scripts=false) but includes clear CLI usage patterns and parameters
References: none bundled (has_references=false)
Instructions: step-by-step guidance for generating diffs (git show, git diff), recommended CLI flags, background execution notes, and post-review validation + remediation workflow.

Compatible agents

Best for agents that can execute shell commands and manage background jobs (CLI-capable assistants like Claude Code, local automation scripts, or CI-integrated bots).

Audit Summary

A code review skill for GLM-5 (744B model) targeting Korean-speaking developers using Claude Code. The SKILL.md is well-structured with clear step-by-step instructions for generating diffs and running reviews. No bundled scripts to test. Major concern: instructs running with dangerouslyDisableSandbox=true, which disables safety sandbox protections. Otherwise clean — no credential leaks, destructive commands, or data exfiltration.

Watch Out

Requires dangerouslyDisableSandbox=true which disables sandbox protections
Written entirely in Korean — limits accessibility
Depends on glm-review CLI being installed and configured

Notes

The dangerouslyDisableSandbox=true instruction is concerning but not malicious — it's needed for the CLI tool to run in background. The skill is well-documented and serves a legitimate purpose. Niche audience due to Korean language and GLM-5 dependency.

Information

Repository: dgk-claude
Stars: 16
Installs: 0

Trust Score

Overall76

Security82

Code Quality72

Architecture68

Usefulness42

More from dgk-claude

/rr — Z.AI Code Review

Run an automated, background code review using a Z.AI (GLM) model to surface security issues, style problems, and PR diffs for staged or uncommitted changes.

Related Skills

Development Worktree

Create an isolated git worktree for feature work, auto-run project setup, and verify a clean test baseline before development.

Readwise Reader Document Management

Manage Readwise Reader documents: list, save, search, move, tag, highlight, export and bulk-edit via official and custom CLIs.

Bounty Hunter — Atlas

Persona skill: 'Atlas' — a profit-focused developer persona for discovering, evaluating and executing paid bounties or freelance tasks with ROI-aware workflows.

Junshi — Research Advisor

Daily strategic research advisor that scans arXiv/venues, digests papers, and proposes bold, ranked research ideas tailored to the user's profile.

Full Stack Builder

End-to-end builder that scaffolds, implements, tests, and optionally deploys web and API applications from a natural-language specification.

ezBookkeeping API Tools

Command-line API tools for ezBookkeeping: record and query transactions, retrieve accounts/categories/tags, and fetch exchange rates for self-hosted personal fi

Feishu Voice Sender

Convert MP3s and send them as native Feishu voice messages (playable voice clips) to users or groups.

Claw Bench

Benchmarking skill that guides an agent through a structured suite of capability tests and reporting steps for leaderboard submission.

Back to Skills