@chachamaru127/harness-mem
v0.19.0
Published
Memory bridge for Claude Code and Codex — local-first, zero-cost coding memory runtime
Maintainers
Readme
Why Harness?
Claude Code is powerful—but sometimes it needs structure.
| Without Harness | With Harness |
|-----------------|--------------|
| Jumps into code immediately | Plans first, then executes |
| Reviews only when asked | Auto-reviews every change |
| Forgets past decisions | SSOT files preserve context |
| rm -rf runs without warning | Dangerous commands blocked |
| One task at a time | Parallel workers |
Three commands. One workflow. Production-ready code.
graph LR
A[Your Idea] --> B["/plan-with-agent"]
B --> C["Plans.md"]
C --> D["/work"]
D --> E["Code + Self-Review"]
E --> F["/harness-review"]
F --> G["Ship It"]Requirements
- Claude Code v2.1+ (Install Guide)
- Node.js 18+ (for safety hooks)
Install in 30 Seconds
# Start Claude Code in your project
claude
# Add the marketplace & install
/plugin marketplace add Chachamaru127/claude-code-harness
/plugin install claude-code-harness@claude-code-harness-marketplace
# Initialize your project
/harness-initThat's it. Start with /plan-with-agent.
Codex CLI Setup
Use Codex CLI with Team Config (shared .codex/):
- Copy
codex/.codexinto your project as.codex - Copy
codex/AGENTS.mdinto your project root asAGENTS.md - Optional: copy
codex/.codex/config.tomland set the MCP server path
Script setup:
/path/to/claude-code-harness/scripts/setup-codex.shClaude Code users can run /setup codex to apply it without leaving the session.
Use $plan-with-agent, $work, $harness-review to run the workflow.
Unified Memory Setup (Claude/Codex/OpenCode shared)
# from your project
/path/to/claude-code-harness/scripts/harness-mem setupThis wires platform adapters, starts harness-memd, and runs smoke + search quality checks.
🪄 TL;DR: Work All
Don't want to read all this? Just type:
/work allOne command. Harness does the rest. Plan → Parallel Implementation → Review → Commit.
graph LR
A["/work all"] --> B["Generate Plan"]
B --> C["Parallel Implementation"]
C --> D["Self-Review"]
D --> E["Quality Gate"]
E --> F["Auto-Commit"]| Before | After |
|--------|-------|
| /plan-with-agent → /work → /harness-review → git commit | /work all |
| 4 commands | 1 |
⚠️ Experimental: Once you approve the plan, Claude runs to completion. Quality gate blocks commit if issues found.
The Core Loop (Details)
1. Plan
/plan-with-agent"I want a login form with email validation"
Harness creates Plans.md with clear acceptance criteria.
2. Work
/work # Auto-detect parallelism
/work --parallel 5 # 5 workers simultaneouslyEach worker implements, self-reviews, and reports.
3. Review
/harness-review| Perspective | Focus | |-------------|-------| | Security | Vulnerabilities, injection, auth | | Performance | Bottlenecks, memory, scaling | | Quality | Patterns, naming, maintainability | | Accessibility | WCAG compliance, screen readers |
Safety First
Harness protects your codebase with hooks:
| Protected | Action |
|-----------|--------|
| .git/, .env, secrets | Write blocked |
| rm -rf, sudo, --force | Confirmation required |
| git status, npm test | Auto-allowed |
| Test tampering | Warning triggered |
45 Skills, Zero Config
Skills auto-load based on context. Use slash commands or natural language.
| Say This | Skill |
|----------|-------|
| "implement login" | impl |
| "review this code" | harness-review |
| "fix the build error" | verify |
| "add Stripe payments" | auth |
| "deploy to Vercel" | deploy |
Key Commands
| Command | What It Does |
|---------|--------------|
| /plan-with-agent | Ideas → Plans.md |
| /work | Execute tasks in parallel |
| /harness-review | 4-perspective review |
| /harness-init | Initialize project |
| /sync-status | Check progress |
| /memory | Manage SSOT files |
Who Is This For?
| You Are | Harness Helps You | |---------|-------------------| | Developer | Ship faster with built-in QA | | Freelancer | Deliver review reports to clients | | Indie Hacker | Move fast without breaking things | | VibeCoder | Build apps with natural language | | Team Lead | Enforce standards across projects |
Architecture
claude-code-harness/
├── skills/ # 45 skill definitions
├── agents/ # 8 sub-agents (parallel workers)
├── hooks/ # Safety & automation
├── scripts/ # Guard scripts
└── templates/ # Generation templatesAdvanced Features
Delegate implementation tasks to OpenAI Codex in parallel:
/work --codex implement these 5 API endpointsCodex implements → Self-reviews → Reports back. Works alongside Claude Code workers.
Setup required: Install Codex CLI and configure API key.
Use Cursor as PM, Claude Code as implementer.
/handoff # Report to Cursor PMPlans.md syncs between both.
Add OpenAI Codex for second opinions:
/harness-review # 4 perspectives + CodexCodex selects 4 relevant experts from 16 specialist types.
Generate product videos with JSON Schema-driven pipeline:
/generate-video- JSON Schema as SSOT (Single Source of Truth)
- 3-layer validation: scene → scenario → E2E
- Remotion-based rendering with deterministic output
Dependencies: Requires Remotion project setup and ffmpeg.
Automatically tracks AI-generated code edits:
.claude/state/agent-trace.jsonl- Records every Edit/Write operation
- Shows project name, current task, recent edits at session end
- Enables
/sync-statusto compare Plans.md with actual changes
No setup required—enabled by default.
Troubleshooting
| Issue | Solution |
|-------|----------|
| Command not found | Run /harness-init first |
| Plugin not loading | Clear cache: rm -rf ~/.claude/plugins/cache/claude-code-harness-marketplace/ and restart |
| Hooks not working | Ensure Node.js 18+ is installed |
For more help, open an issue.
Uninstall
/plugin uninstall claude-code-harnessProject files (Plans.md, SSOT files) remain unchanged.
Documentation
| Resource | Description | |----------|-------------| | Changelog | Version history | | Claude Code Compatibility | Requirements | | Cursor Integration | 2-Agent setup |
Contributing
Issues and PRs welcome. See CONTRIBUTING.md.
Acknowledgments
License
MIT License — Free to use, modify, commercialize.
