⚠
High Risk:This skill has significant security concerns. Review the findings below before installing.

glm-coding-agent

⚠Caution·Scanned 2/17/2026

High-risk skill: runs local shell commands (including bash -c "$(curl -fsSL https://raw.githubusercontent.com/openclaw/openclaw/main/skills/glm-coding-agent/install.sh)") and invokes --dangerously-skip-permissions. It reads "${HOME}/.openclaw/openclaw.json" to export ANTHROPIC_AUTH_TOKEN and calls https://api.z.ai/api/anthropic.

from clawhub.ai·v3e835f6·16.5 KB·0 installs
Scanned from 2.0.0 at 3e835f6 · Transparency log ↗
$ vett add clawhub.ai/cgnl/glm-coding-agentReview security findings before installing

GLM Coding Agent

Free coding assistant using Claude Code CLI with GLM 4.7 (200k context, $0 cost).

Quick Start

# One command setup
bash -c "$(curl -fsSL https://raw.githubusercontent.com/openclaw/openclaw/main/skills/glm-coding-agent/install.sh)"

# Or manual:
mkdir -p ~/clawd/scripts
cat > ~/clawd/scripts/glmcode.sh << 'EOF'
#!/bin/bash
export ANTHROPIC_BASE_URL="https://api.z.ai/api/anthropic"
export ANTHROPIC_AUTH_TOKEN="YOUR_TOKEN_HERE"
export API_TIMEOUT_MS=3000000
exec claude --settings ~/.claude/settings-glm.json "$@"
EOF
chmod +x ~/clawd/scripts/glmcode.sh

Usage

# From OpenClaw
bash pty:true workdir:~/project command:"~/clawd/scripts/glmcode.sh 'Your task'"

# Or via sub-agent
sessions_spawn({ task: "Build todo API", model: "zai/glm-4.7" })

See SKILL.md for full docs.

Why?

  • ✅ Free (0 cost)
  • ✅ 200k context
  • ✅ Fast
  • ✅ Works with Claude Code CLI

Perfect for refactoring, bug fixes, docs, and simple features!