Define AI workflows like Makefile targets.
Framework verbs are fixed. Project skills are replaceable.
AI coding agents skip diagnosis, forget reviews, and repeat mistakes. Summoner adds a process layer — checkpoints between phases, post-game reviews, and a memory chain that auto-recalls past fixes.
In Claude Code:
/plugin marketplace add johnson-xue/summoner
/plugin install summoner@summoner-marketplace
Then in your project (hooks work immediately — pre-compiled for macOS/Linux):
cd your-project
~/.claude/plugins/summoner/scripts/summoner-init.sh
~/.claude/plugins/summoner/scripts/init-memory-db.sh $(grep -A2 '^project:' summoner.yaml | grep 'name:' | head -1 | awk '{print $2}')Alternative:
git clone https://github.com/johnson-xue/summoner.git ~/.claude/plugins/summoner/— same result, no marketplace needed. If on an unsupported platform:cd ~/.claude/plugins/summoner/hooks && make build.
/summoner:fix "线上报错 SC_ErrInnerLogic..."
Phase 0 Memory — auto-recalls past bug patterns
Phase 1 Diagnosis — root cause (iron law: never skip)
⏸️ Checkpoint → [enter] continue [skip] skip [recall] redo [stop]
Phase 2 Reproduce — Prove-It test (auto-skipped for config-only fixes)
Phase 3 Fix — apply the fix
Phase 4 Verify — run test suite
Phase 5 Review — code review
📋 Post-Game Review — 5-type questionnaire, auto-journaled to SQLite
| Command | Does |
|---|---|
/summoner:fix |
Bug fix: diagnose → reproduce → fix → verify → review |
/summoner:new |
Feature: define → plan → implement → test → review |
/summoner:ship |
Pre-launch: fan-out 1-3 reviewers → merge → go/no-go |
/summoner:debug |
Diagnosis only — no code changes |
/summoner:ops |
Server operations (delegated to project skill) |
/summoner:review |
Standalone code review |
| Commands | Memory | Hooks | |
|---|---|---|---|
| Claude Code | ✅ | ✅ SQLite | ✅ Go |
| Gemini CLI | ✅ | ✅ bash | — |
| OpenCode | ✅ | ✅ bash | — |
| Cursor / Windsurf / Copilot / Aider | ✅ | ✅ bash | — |
| Workflow | Tokens | vs Direct |
|---|---|---|
/summoner:fix (with memory) |
~9,300 | +86% |
/summoner:fix (no memory hit) |
~8,300 | +66% |
/summoner:debug (diagnose only) |
~4,300 | +35% |
Multi-step workflows → Summoner. Single-step tasks → direct skill.
63 files · 19 core · compiled hooks ~7.5 MB · zero external deps beyond Go + SQLite3
Summoner is in early development (v0.1.1). Your feedback shapes its direction.
| What | Where |
|---|---|
| Bug report | GitHub Issues — include version, platform, and what you expected |
| Feature request | GitHub Issues — tell us your use case and current workaround |
| Questions / Ideas | GitHub Discussions — Q&A, ideas, show your setup |
| Known Issues | See v0.1.1 Release Notes for current limitations |
Why report? Every issue you file teaches Summoner what real-world AI workflows need. Even a single sentence helps — "I tried X, expected Y, got Z."
- anthropics/skills — official skill examples
- obra/superpowers — general-purpose Claude Code skills
- addyosmani/agent-skills — production engineering skills
MIT © Jingshan Xue