A collection of OpenCode skills I've built from real use — things I reach for daily and have refined through trial and error. If something doesn't work for you, or you see a gap, open an issue or send a PR.
Current version: 0.9.0 — see CHANGELOG
# List available skills
npx skills add mskadu/opencode-agent-skills --list
# Install a specific skill to OpenCode
npx skills add mskadu/opencode-agent-skills --skill permission-manager --agent opencode
# Install all skills globally
npx skills add mskadu/opencode-agent-skills --all -g -ycurl -fsSL https://raw.githubusercontent.com/mskadu/opencode-agent-skills/main/install.sh | bashgit clone https://github.com/mskadu/opencode-agent-skills.git
cp -r opencode-agent-skills/skills/* ~/.config/opencode/skills/| Skill | Description | Risk |
|---|---|---|
| anti-sycophancy | Eliminate sycophantic agreement patterns in AI responses | safe |
| permission-manager | Review, configure, and audit OpenCode permission settings | critical |
| skill-suggester | Scan prompt history for recurring patterns, propose new skills | safe |
| ollama-sync | Synchronize Ollama model list with OpenCode configuration | safe |
| smart-git-automation | Smart change detection, auto branch naming, streamlined commit/PR | critical |
- OpenCode CLI
MIT