The performance optimization system for Kilo CLI.
Not just configs. A complete system: 30 agents, 136 skills, 60 commands, rules, and MCP configurations evolved over 10+ months of intensive daily use. Originally from everything-claude-code, ported and adapted for Kilo CLI.
Installs to ~/.config/kilo/ for use across all projects:
git clone https://github.com/DoozieSoftware/everything-kilo-code.git
cd everything-kilo-code
npm install
./install.sh --target kilo --profile fullInstalls to ./.kilo/ in the current project only:
git clone https://github.com/DoozieSoftware/everything-kilo-code.git
cd everything-kilo-code
npm install
./install.sh --target kilo --profile full --project./install.sh --target kilo typescript php pythonThis installs everything to ~/.config/kilo/:
| Component | Location | Count |
|---|---|---|
| Agents | ~/.config/kilo/agent/*.md |
30 |
| Commands | ~/.config/kilo/command/*.md |
60 |
| Skills | ~/.config/kilo/skill/*/SKILL.md |
136 |
| Rules | ~/.config/kilo/rules/{common,typescript,...}/ |
77 |
| MCP Config | ~/.config/kilo/kilo.jsonc |
19 servers |
Specialized subagents for delegation:
| Agent | Purpose |
|---|---|
planner |
Feature planning, architectural changes |
architect |
System design decisions |
tdd-guide |
Test-driven development guidance |
code-reviewer |
Code quality and security review |
security-reviewer |
Vulnerability analysis |
build-error-resolver |
Fix build/compile errors |
e2e-runner |
Playwright E2E testing |
refactor-cleaner |
Dead code cleanup |
doc-updater |
Documentation sync |
docs-lookup |
API reference research |
database-reviewer |
Database/SQL review |
typescript-reviewer |
TypeScript/JavaScript review |
python-reviewer |
Python code review |
go-reviewer / go-build-resolver |
Go review and build fixes |
java-reviewer / java-build-resolver |
Java/Spring Boot |
kotlin-reviewer / kotlin-build-resolver |
Kotlin/Android |
rust-reviewer / rust-build-resolver |
Rust |
cpp-reviewer / cpp-build-resolver |
C++ |
pytorch-build-resolver |
PyTorch/CUDA training errors |
loop-operator |
Autonomous loop execution |
harness-optimizer |
Harness config tuning |
Slash commands for quick actions:
Planning & Quality:
/plan, /tdd, /code-review, /build-fix, /refactor-clean, /verify, /checkpoint, /test-coverage, /e2e, /eval
Language-Specific:
/go-review, /go-test, /go-build, /python-review, /cpp-review, /cpp-build, /cpp-test, /kotlin-review, /kotlin-build, /kotlin-test, /rust-review, /rust-build, /rust-test
Learning & Memory:
/learn, /learn-eval, /evolve, /instinct-status, /instinct-import, /instinct-export, /prune
Session Management:
/sessions, /save-session, /resume-session
Documentation:
/update-docs, /update-codemaps, /docs
Multi-Agent:
/multi-plan, /multi-execute, /multi-backend, /multi-frontend, /multi-workflow, /orchestrate
Harness:
/harness-audit, /quality-gate, /model-route, /context-budget, /prompt-optimize, /loop-start, /loop-status
Auto-activating domain workflows organized by category:
Core: tdd-workflow, security-review, coding-standards, continuous-learning, verification-loop, eval-harness, search-first, autonomous-loops
Languages: typescript-*, python-*, golang-*, laravel-*, django-*, springboot-*, cpp-*, perl-*, swift-*, kotlin-*, rust-*, nuxt4-*, flutter-*
DevOps: deployment-patterns, docker-patterns, database-migrations, api-design, e2e-testing, postgres-patterns
Business: article-writing, content-engine, market-research, investor-materials, investor-outreach
AI/LLM: cost-aware-llm-pipeline, content-hash-cache-pattern, regex-vs-llm-structured-text, prompt-optimizer
3 enabled by default:
sequential-thinking— Chain-of-thought reasoningmemory— Persistent memory across sessionscontext7— Live documentation lookup
16 available (disabled, enable as needed): github, firecrawl, supabase, playwright, exa-web-search, vercel, railway, cloudflare-docs, clickhouse, magic, filesystem, fal-ai, browserbase, token-optimizer, confluence
---
description: Expert planning specialist for complex features.
mode: subagent
model: anthropic/claude-opus
hidden: false
---
System prompt for this agent.---
description: Create implementation plan.
---
Command body with $ARGUMENTS support.---
name: tdd-workflow
description: Use this skill when writing new features.
---
Skill body with workflow instructions.@planner Add user authentication
/tdd
/code-review src/services/
/build-fix
/plan
/verify
/learn
everything-kilo-code/
├── .kilo/ # Kilo CLI native directory
│ ├── agent/ # 30 agent definitions (Kilo format)
│ ├── command/ # 60 slash commands (Kilo format)
│ ├── skill/ # 136 skills (Kilo format)
│ └── rules/ # Rule files (common, typescript, php, ...)
├── agents/ # Source agent files (ECC format)
├── commands/ # Source command files (ECC format)
├── skills/ # Source skill files (ECC format)
├── rules/ # Source rule files
├── mcp-configs/ # MCP server definitions
├── manifests/ # Install module definitions
├── scripts/ # Install and utility scripts
├── kilo.jsonc # Kilo config template
├── AGENTS.md # Agent instructions
├── KILO.md # Project context for Kilo CLI
├── install.sh # Installer (Linux/macOS)
└── install.ps1 # Installer (Windows)
cd everything-kilo-code
git pull origin main
./install.sh --target kilo --profile fullEKC is forked from everything-claude-code. When the upstream repo publishes new agents, skills, or commands, you can sync them into EKC. See SYNC.md for the full migration guide.
Quick sync:
# Add upstream remote (once)
git remote add upstream https://github.com/affaan-m/everything-claude-code.git
# Fetch and review upstream changes
git fetch upstream
git log --oneline upstream/main..HEAD
# Cherry-pick or merge specific upstream commits
git cherry-pick <commit-hash>
# Convert new components to Kilo format
node scripts/convert-to-kilo.js
# Reinstall
./install.sh --target kilo --profile fullForked from everything-claude-code by Affaan Mustafa. Adapted for Kilo CLI by DoozieSoftware.
MIT — Use freely, modify as needed, contribute back if you can.
{ "instructions": ["rules/common/**/*.md", "rules/typescript/**/*.md"], "skills": { "paths": [".kilo/skill"] }, "mcp": { "sequential-thinking": { "type": "local", "command": [...], "enabled": true } } }