\n\n**Kodelyth ECC** is a production-grade AI coding toolkit — **70 specialist agents (incl. an 8-agent devil-mode adversarial crew), 196 skills, 103 commands**, a god-tier **semantic intent-routing system**, local self-learning memory, MCP server, swarm orchestrator, and an observability dashboard — all local, zero telemetry.\n\nNow bundled with:\n\n- **RTK** (Rust Token Killer, Apache-2.0) — auto-compresses shell command output for **60-90% input token savings**\n- **Terse mode** (ECC-native, inspired by Caveman) — 4-level output-token compression dial for **40-70% output savings**\n- **Codebase graph** (via [codebase-memory-mcp](https://github.com/DeusData/codebase-memory-mcp), MIT) — AST-parsed knowledge graph across 158 languages; structural queries at **99% fewer tokens** than file-by-file grep\n- **Interactive CLI** — type `kodelythecc` alone in a terminal for an arrow-key menu with live update check, dashboard, background daemon\n\nWorks with **Claude Code**, **Windsurf**, **Cursor**, **Codex CLI**, **Google Antigravity**, **OpenCode**, **Cline**, **Roo Code**, **Aider**, **Kimi**, and **Gemini CLI**.\n\n> No telemetry. No cloud. Just rules, agents, skills, MCP server, and your own private memory store — all on your disk. The local dashboard gives you full visibility without sending anything anywhere.\n\n---\n\n## Why ECC ≠ Another Agent Collection\n\nMost \"AI agent kits\" are folders of markdown files you have to remember the names of. **ECC is infrastructure** — a layered system where intent routing, compound memory, parallel orchestration, and quality hooks all reinforce each other.\n\n```\nYou: \"I've been staring at this NullPointerException for two hours,\n I'm losing my mind.\"\n\nAI: → Routing to debug-detective (your error + frustration matches the bug-tracking signal)\n\n That kind of bug is exhausting — let's trace it properly so we\n fix the root cause, not the symptom.\n\n First, can you share the full stack trace and...\n```\n\nYou never typed `use debug-detective`. You didn't have to. The toolkit read the intent, picked the specialist, and announced the routing. Next time you can invoke it directly — but you don't have to remember names to get senior-grade help.\n\n### The Layer Stack\n\n| Layer | What it does | Other kits |\n|---|---|---|\n| **Intent routing** | Plain-language → right specialist via 10-tier priority rules | Mostly missing — you memorize names |\n| **70 agents** | Specialists with playbooks, severity calibration, real commands | Often persona-only (\"you are a senior engineer...\") |\n| **196 skills** | Domain knowledge files agents read on demand | Rarely separated from agents |\n| **103 commands** | Slash workflows (`/tdd`, `/arena`, `/devil-mode`, `/team-review`) | Limited or none |\n| **8 parallel commands** | Fire 3-8 agents simultaneously, aggregate results | Rare |\n| **Compound memory** | BM25 local recall + auto-inject + project lessons | Cloud-only or absent |\n| **22+ hooks** | Quality gates, secret scan, project-DNA detection | Often missing |\n| **11 IDE platforms** | Claude Code, Windsurf, Cursor, Codex, Antigravity, OpenCode, Cline, Roo Code, Aider, Kimi, Gemini CLI (13 install targets) | 1-2 platforms typical |\n| **Zero telemetry** | Everything stays on your disk; verifiable | Many kits phone home |\n\n### Quick Comparison vs Other Kits\n\n| Feature | **Kodelyth ECC** | `agency-agents` | `awesome-claude-agents` | Generic prompt libs |\n|---|---|---|---|---|\n| Specialist agents | **70** | ~30 | ~20 | Varies |\n| Skills as separate layer | ✅ 194 | ❌ | ❌ | ❌ |\n| Slash commands | ✅ 97 | Some | Some | ❌ |\n| **Parallel multi-agent commands** | ✅ 8 (incl. `/devil-mode`) | ❌ | ❌ | ❌ |\n| **Intent routing (plain language → agent)** | ✅ 10-tier rule | ❌ | ❌ | ❌ |\n| **Local BM25 self-learning memory** | ✅ | ❌ | ❌ | ❌ |\n| **Compound learning from corrections** | ✅ `tasks/lessons.md` | ❌ | ❌ | ❌ |\n| **Adversarial / red-team agents** | ✅ 8 (devil-mode) | ❌ | ❌ | ❌ |\n| **Adversarial build/attack loop** | ✅ `/arena` — scored, verified, converges | ❌ | ❌ | ❌ |\n| Quality hooks | ✅ 22+ | Some | ❌ | ❌ |\n| IDE platforms | **11** (Claude Code, Windsurf, Cursor, Codex, Antigravity, OpenCode, Cline, Roo Code, Aider, Kimi, Gemini CLI) | 1-2 | 1 | Varies |\n| Telemetry | ❌ none | Varies | ❌ | Varies |\n| Test coverage | ✅ 373 tests | ❌ | ❌ | ❌ |\n| Distributed via | `npx`, curl, clone | Manual | Manual | Manual |\n\n---\n\n## Install\n\n
\n\n
\n\n### The lazy install (one command, everything wired up)\n\n```bash\nnpm i -g kodelyth-ecc\nkodelythecc --target claude-code --codebase-graph\n# then restart your AI tool\n```\n\nThat's it. This single flow:\n\n1. Installs both binaries (`kodelyth-ecc` and short-form `kodelythecc`) to your PATH\n2. Copies 70 agents + 196 skills + 103 commands + 22 hooks + 15 rules into your AI tool's config dir\n3. Auto-installs **RTK** binary and wires its PreToolUse hook (input compression starts on next AI restart)\n4. Installs **Terse mode** skill + `/terse` and `/terse-compress` slash commands (dormant — user types `/terse` to activate)\n5. Auto-installs **codebase-memory-mcp** and registers its MCP entries in your AI tool (with `--codebase-graph`)\n6. Runs legacy-memory migration if you had `~/.kodelyth/` from an older install → `~/.kodelythecc/`\n\nAfter it finishes, run `kodelythecc` alone to open the interactive menu.\n\n### Option 1 — npx from npm (recommended, any platform)\n\n```bash\nnpx kodelyth-ecc # Claude Code (default)\nnpx kodelyth-ecc --target windsurf-project # Windsurf (per-project)\nnpx kodelyth-ecc --target windsurf-home # Windsurf (global)\nnpx kodelyth-ecc --target cursor-project # Cursor IDE\nnpx kodelyth-ecc --target codex-home # Codex CLI\nnpx kodelyth-ecc --target antigravity # Google Antigravity\nnpx kodelyth-ecc --target opencode # OpenCode\nnpx kodelyth-ecc --target cline # Cline (VS Code)\nnpx kodelyth-ecc --target roocode # Roo Code (VS Code)\nnpx kodelyth-ecc --target aider # Aider terminal agent\nnpx kodelyth-ecc --target kimi # Kimi Code\nnpx kodelyth-ecc --target gemini-project # Gemini CLI (project)\nnpx kodelyth-ecc --target gemini-home # Gemini CLI (global)\n```\n\nNode.js 18+ required. [Download Node](https://nodejs.org) if you don't have it.\n\n### Platform support at a glance\n\nFeature depth varies by platform — hooks are a Claude Code native format, and some platforms have no agent/command concept:\n\n| Platform | Agents | Skills | Commands | Hooks | Rules |\n|---|---|---|---|---|---|\n| **Claude Code** | ✓ 70 | ✓ 194 | ✓ 97 | ✓ 22+ | ✓ |\n| **Roo Code** | ✓ | ✓ | ✓ | — | ✓ |\n| **Codex CLI** | ✓ | ✓ | ✓ | — | ✓ |\n| **Aider** | ✓ | ✓ | ✓ | — | ✓ |\n| **Kimi** | ✓ | ✓ | ✓ | — | ✓ |\n| **Windsurf** | ✓ | ✓ | — | — | ✓ |\n| **Antigravity** | ✓ | partial | ✓ | — | ✓ |\n| **Gemini CLI** | ✓ | ✓ | — | — | ✓ |\n| **Cursor** | — | ✓ | — | — | ✓ |\n| **Cline** | ✓ | — | ✓ | — | ✓ |\n| **OpenCode** | — | — | — | — | ✓ |\n\nHooks use Claude Code's JSON settings format — no equivalent exists on other platforms. Cursor reads rules and skills from `.cursor/`; its agent system uses a different format not yet compatible with ECC agents.\n\n### Memory & dashboard reality per IDE\n\nMemory storage is a single shared file at `~/.kodelyth/memory/memories.jsonl` — every IDE on the same machine reads and writes the **same** memories. The only thing that varies is *how* memories surface:\n\n| Platform | Auto-recall on every prompt | Auto-capture on success | Manual recall via MCP tool | Dashboard \"Live IDE activity\" |\n|---|---|---|---|---|\n| **Claude Code** | ✓ (hook) | ✓ (hook) | ✓ | ✓ Claude session files |\n| **Windsurf** | — | — | ✓ `recall_memory` | ✓ Windsurf + Windsurf-Next state |\n| **Cursor** | — | — | ✓ `recall_memory` | ✓ workspace storage dirs |\n| **Codex CLI** | — | — | ✓ `recall_memory` | — |\n| **Antigravity** | — | — | ✓ `recall_memory` | ✓ `.agent/` in cwd |\n| **Roo Code / Aider / Kimi / Cline / Gemini CLI / OpenCode** | — | — | ✓ if MCP-capable | — |\n\nWhat this means in practice:\n- A memory captured in **Claude Code** is recall-able from every other IDE the same day — the file is shared.\n- In **Windsurf / Cursor / Codex / Antigravity**, the AI does NOT auto-fire memory recall; the `rules/common/memory-protocol.md` rule (installed automatically) tells the AI to call the `recall_memory` MCP tool proactively at the start of substantive prompts.\n- The dashboard's **Sessions → Live IDE activity** tab surfaces session files for Claude Code, Windsurf, Windsurf-Next, Cursor, and Antigravity. Add custom paths via the `KODELYTH_EXTRA_IDE_WATCH` env var (comma-separated).\n\n```bash\n# Watch additional paths in the dashboard\nexport KODELYTH_EXTRA_IDE_WATCH=\"$HOME/my-agent-logs,$HOME/other-tool/state\"\nnpx kodelyth-ecc dashboard\n```\n\n### Option 2 — npx from GitHub (always latest commit)\n\n```bash\nnpx github:sifxprime/kodelyth-ecc\n```\n\nSame `--target` flags work.\n\n### Option 3 — curl (macOS / Linux only)\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/sifxprime/kodelyth-ecc/main/install.sh | bash\n```\n\nWith a target:\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/sifxprime/kodelyth-ecc/main/install.sh | bash -s -- --target windsurf-project\n```\n\n### Power Bundles\n\nPre-configured for who you actually are:\n\n```bash\nnpx kodelyth-ecc --bundle indie-hacker # Solo founder / SaaS — ship fast, validate, harden\nnpx kodelyth-ecc --bundle red-team # Security engineer — devil-mode + adversarial workflows\nnpx kodelyth-ecc --bundle enterprise # Compliance / audit team — SBOM, license, supply chain\n```\n\nEach bundle installs the full ECC toolkit (all 70 agents, 196 skills, 103 commands, 22+ hooks), adds a `BUNDLE.md` cheat sheet, and biases the AI toward audience-fit workflows on every session.\n\nCombine with any target:\n\n```bash\nnpx kodelyth-ecc --bundle red-team --target windsurf-project\nnpx kodelyth-ecc --bundle enterprise --target codex-home\nnpx kodelyth-ecc --bundle indie-hacker --target antigravity\n```\n\n### Option 4 — Clone and run\n\n```bash\ngit clone https://github.com/sifxprime/kodelyth-ecc.git\ncd kodelyth-ecc\n\n# macOS / Linux\n./install.sh # Claude Code (default)\n./install.sh --target windsurf-project # Windsurf\n\n# Windows (PowerShell)\n.\\install.ps1\n.\\install.ps1 -Target windsurf-project\n```\n\n---\n\n## MCP Server — Universal Adapter\n\n
\n\n
\n\nRun ECC as a Model Context Protocol server and consume it from Claude Desktop, LangGraph, AutoGen, CrewAI, OpenAI Agents SDK, Cursor, Windsurf — anything that speaks MCP.\n\n```bash\nnpx kodelyth-ecc mcp # stdio JSON-RPC server\n```\n\nWhat it exposes (all local, zero telemetry):\n\n- **16 tools** — `route_intent`, `recall_memory`, `capture_memory`, `list_agents`, `get_skill`, `audit_skill_match`, …\n- **6 prompts** — full intent routing rule, agents/skills/commands overviews, handoff chains, devil-mode\n- **365 resources** — every agent, skill, command, rule, and bundle addressable via `kodelyth://...` URIs\n\nDrop into Claude Desktop in 30 seconds:\n\n```jsonc\n// claude_desktop_config.json\n{\n \"mcpServers\": {\n \"kodelyth-ecc\": {\n \"command\": \"npx\",\n \"args\": [\"-y\", \"kodelyth-ecc\", \"mcp\"]\n }\n }\n}\n```\n\nFull reference: [`docs/mcp.md`](docs/mcp.md).\n\n---\n\n## How Intent Routing Works\n\n
\n\n
\n\nThe toolkit ships with a single rule file (`rules/common/agent-intent-routing.md`) that the AI loads automatically on every session. It maps **what you say** to **the right specialist agent** across 10 priority tiers.\n\n### Two activation paths\n\n**1. Explicit** — type it directly:\n```\nuse debug-detective\n@code-reviewer\ninvoke security-reviewer\n```\n\n**2. Implicit** — just describe your problem; the AI routes you:\n\n| What you write | Auto-routed to |\n|---|---|\n| \"I'm stuck, no idea where to start\" | `kodelyth-advisor` |\n| \"I've been debugging this for hours\" | `debug-detective` |\n| \"nothing works, driving me crazy\" | `debug-detective` |\n| \"Should I use Context or Zustand?\" | `pair-programmer` |\n| \"help me build a todo app\" | `/project-launch` |\n| \"I have this idea for a SaaS dashboard\" | `/project-launch` |\n| \"I'm starting a new side project\" | `/project-launch` |\n| \"can you review my code?\" | `code-reviewer` or `/team-review` |\n| \"review my project before I deploy\" | `/team-review` |\n| \"is my project ready to ship?\" | `/team-review` |\n| \"my site looks plain, needs visuals\" | `image-architect` |\n| \"I need an OG image for my app\" | `image-architect` |\n| \"remember we always use pnpm here\" | `/lessons` |\n| \"Build failed on Vercel\" | `build-error-resolver` |\n| \"Is this JWT signing secure?\" | `security-reviewer` |\n| \"Why is this so slow?\" | `performance-optimizer` |\n| \"Plan the v2 migration\" | `planner` + `migration-guide` |\n| \"Tests pass locally but fail on CI\" | `flake-hunter` + `env-debugger` |\n| \"I lost my commits after `reset --hard`\" | `git-rescue` |\n| \"npm install is failing\" | `dependency-doctor` |\n| \"Cut a 1.4 release\" | `release-captain` |\n| \"Add accessibility to this form\" | `ux-reviewer` |\n| \"Open-source this project\" | `opensource-forker` (chain) |\n| [paste code with no text] | `code-reviewer` |\n| [paste stack trace with no text] | `debug-detective` |\n\nThe AI **always announces** which agent is taking over (`→ Routing to `) and **always teaches** you the explicit form for next time (`Tip: type \"use \"`). No silent personality changes.\n\n### Multi-agent chains\n\nReal problems span multiple specialties. ECC ships standard handoff chains:\n\n```\npair-programmer → tdd-guide → code-reviewer → security-reviewer\n(approach) (write tests) (review impl) (auth, validation)\n```\n\n```\ndebug-detective → tdd-guide → refactor-cleaner\n(root cause) (regression test) (cleanup)\n```\n\n```\nopensource-forker → opensource-sanitizer → opensource-packager → release-captain\n(clean fork) (strip secrets) (README, license) (cut v0.1.0)\n```\n\nSee `skills/agent-handoff/SKILL.md` for the full handoff protocol and standard chains.\n\n---\n\n## Parallel Agents — 8 Commands, Minutes Not Hours\n\n
\n\n
\n\nEight commands fire multiple specialist agents simultaneously and aggregate their results into a single structured report.\n\n| Command | Agents Fired | Time Saved |\n|---|---|---|\n| `/project-launch` | architect + pair-programmer + security-reviewer + tdd-guide + ux-reviewer | 45 min → 10 min |\n| `/team-review` | code-reviewer + security-reviewer + performance-optimizer + api-guardian | 60 min → 15 min |\n| `/security-audit` | security-reviewer + dependency-doctor + api-guardian | 30 min → 8 min |\n| `/debug-blitz` | debug-detective + silent-failure-hunter + env-debugger | 60 min → 15 min |\n| `/refactor-sprint` | refactor-cleaner + code-simplifier + type-design-analyzer + tdd-guide | 45 min → 12 min |\n| `/pre-release` | release-captain + security-reviewer + code-reviewer | 30 min → 8 min |\n| `/onboard` | code-explorer + architect + doc-updater | 45 min → 12 min |\n| `/devil-mode` | 8 adversarial agents (see below) | Hours → 20 min |\n| `/arena` | GOD crew vs EVIL crew, looped until the attacker gives up | Days → 1 session |\n\nEach command waits for all agents to complete, then returns a single **Team Review Report** with findings bucketed by severity: CRITICAL → HIGH → MEDIUM → LOW.\n\n---\n\n## Devil Mode — 8 Adversarial Agents\n\n
\n\n
\n\n8 adversarial agents that read your codebase the way an attacker would. Fire them in parallel with `/devil-mode`:\n\n```\n/devil-mode --pre-public # before going open-source — full secret/license/IP sweep\n/devil-mode --pre-launch # before launch — adds AI red-team + chaos planning\n/devil-mode --all # all 8 adversarial agents in parallel\n```\n\nThe crew: `prompt-injection-hunter`, `supply-chain-auditor`, `secret-hunter`, `license-violation-finder`, `jailbreak-tester`, `code-stealer-detector`, `backdoor-hunter`, `chaos-engineer`. Each ships with real bash-grep patterns, severity calibration, and remediation playbooks — not theatrical personas.\n\n---\n\n## Live Dashboard — Full Visibility, Localhost Only\n\n
\n\n
\n\n**You have a working, real-time dashboard.** One command launches it:\n\n```bash\nnpx kodelyth-ecc dashboard\n# Opens http://127.0.0.1:5747 in your browser\n```\n\nWhat's inside:\n\n| Tab | What you see |\n|---|---|\n| **Overview** | Agent count, memory stats, session count, recent activity |\n| **Memory** | Browse, search, and manage your local BM25 memory store |\n| **Evolve** | Self-improving memory — review AI-proposed refinements |\n| **Catalog** | Full searchable index of all 70 agents, 196 skills, 103 commands |\n| **Sessions** | **Live IDE activity** (Claude Code, Windsurf, Windsurf-Next, Cursor, Antigravity) + orchestration/swarm sessions |\n\nReal-time:\n- SSE push every 3 seconds, only when at least one browser tab is connected (zero CPU otherwise)\n- Watches memory writes, evolve signals, token-budget changes, and IDE session files\n- \"Last activity\" auto-updates without a page refresh\n- Set `KODELYTH_EXTRA_IDE_WATCH=path1,path2` to watch additional paths\n\nSecurity design:\n- **GET-only** — no write endpoints accessible from the UI\n- **Localhost-bound** — refuses connections from non-localhost Host headers (DNS rebinding protection)\n- **Hardened headers** — `Content-Security-Policy`, `X-Frame-Options`, `X-Content-Type-Options` on every response\n- **Zero telemetry** — no external network calls ever leave your machine\n- **Max 10 SSE clients** — connection cap prevents resource exhaustion\n\nFull reference: [`docs/dashboard.md`](docs/dashboard.md).\n\n---\n\n## RTK — Input Token Savings (60-90%)\n\n
\n\n[RTK (Rust Token Killer)](https://github.com/rtk-ai/rtk) — Apache-2.0, single Rust binary — intercepts shell commands before they run and filters the output. Common commands like `git status`, `ls`, `cargo test`, `docker ps` return ~80% less text with zero information loss for the LLM.\n\nECC bundles RTK end-to-end:\n\n- Auto-installs the binary via Homebrew (macOS) or curl script (Linux / WSL)\n- Auto-configures the PreToolUse hook in your AI tool's settings\n- Live ledger surfaced in the dashboard's **Token Savings** tab\n\n### Manage from the CLI\n\n```bash\nkodelythecc rtk install # install rtk binary\nkodelythecc rtk enable --target X # wire into one IDE\nkodelythecc rtk enable --all # wire into every ECC-installed IDE at once\nkodelythecc rtk status # version + active integrations\nkodelythecc rtk gain --all # raw rtk savings output\nkodelythecc rtk --help # focused help\n```\n\n### Live proof (from this repo maintainer's Mac)\n\n```\ntotal_commands: 1,285\ntotal_input: 7,964,612 (raw)\ntotal_output: 2,858,501 (after RTK filter)\ntotal_saved: 5,107,394 ← 64.1% average reduction\n```\n\n---\n\n## Terse Mode — Output Token Savings (40-70%)\n\n
\n\nRTK compresses input. **Terse mode** compresses output. Together they stack — savings on both sides of every turn.\n\nECC's own implementation, inspired by [Caveman](https://github.com/JuliusBrussee/caveman) (MIT), independently written. Ships as a skill + two slash commands + a deterministic zero-dep memory-file compressor.\n\n### Four levels\n\nType `/terse` in your AI tool. Level sticks until you switch or session ends.\n\n| Level | Style |\n|---|---|\n| `/terse off` | Normal AI voice |\n| `/terse lite` | Light trim, drop filler (\"basically\", \"essentially\") |\n| `/terse full` | Telegram-style fragments (default) |\n| `/terse ultra` | Maximum compression, symbols over words |\n\n**Byte-preserved always**: fenced code blocks, inline code, shell commands, error text, URLs, file paths, identifiers, numbers, versions.\n\n### Compress memory files permanently\n\n`/terse-compress` (or `kodelythecc terse compress `) rewrites `CLAUDE.md`-style memory files into terse form so they cost fewer tokens **every session forever**. ~30% average byte reduction on real prose, 100% code/URL/path integrity.\n\n### CLI\n\n```bash\nkodelythecc terse status # skill install state\nkodelythecc terse stats # tokens saved from ledger\nkodelythecc terse compress # rewrite a file, byte-preserves code\nkodelythecc terse enable --all # install skill + commands into every ECC IDE\nkodelythecc terse --help\n```\n\n---\n\n## Codebase Graph — 158 Languages, Structural Queries\n\n
\n\nPowered by [DeusData/codebase-memory-mcp](https://github.com/DeusData/codebase-memory-mcp) (MIT). AST-parsed knowledge graph via tree-sitter across **158 languages**, Hybrid LSP semantic type resolution for 11 major languages, cross-service HTTP/gRPC/GraphQL linking, 14 MCP tools for structural queries.\n\n**Why this matters**: \"Who calls `ProcessOrder`?\" via file-by-file grep = ~412k tokens. Same question via the graph = ~3.4k tokens. That's a **99% reduction** on structural questions.\n\n### Auto-install path\n\nPass `--codebase-graph` on ECC install and it auto-runs their official curl script + auto-configures the MCP server entry in every installed AI-coding agent.\n\n```bash\nkodelythecc --target claude-code --codebase-graph\n# ↑ installs ECC + RTK + Terse + codebase-memory-mcp, all wired up\n```\n\nThen in your AI tool, say **\"Index this project\"**. Done.\n\n### CLI\n\n```bash\nkodelythecc codebase install # install + auto-register\nkodelythecc codebase status # version + indexed projects\nkodelythecc codebase query search_graph '{\"name_pattern\": \".*Handler.*\"}'\nkodelythecc codebase query trace_path '{\"function_name\": \"main\"}'\nkodelythecc codebase query get_architecture '{}'\nkodelythecc codebase --help\n```\n\n---\n\n## Interactive CLI Menu\n\n
\n\nType `kodelythecc` alone in a real terminal → arrow-key menu opens.\n\n```\n⚙ Kodelyth ECC v2.3.0 · Elite Code Crew up to date\n\n ▸ Open Dashboard localhost — RTK, Terse, Codebase, Memory\n Install ECC for another IDE 13-target picker\n RTK status Version + wired IDEs + savings\n Terse status Skill install state, ledger totals\n Codebase graph status 158 languages, structural queries\n Memory stats BM25 recall — captures, projects, tags\n Run in background Detached dashboard daemon\n Exit\n\n↑/↓ navigate · ⏎ select · q / esc / Ctrl+C to quit\n```\n\n### Behaviour rules\n\n| Situation | Menu opens? |\n|---|---|\n| `kodelythecc` in Terminal / iTerm | **Yes** |\n| `kodelythecc rtk status` (any subcommand) | No — runs subcommand |\n| `echo hi \\| kodelythecc` (piped) | No — runs installer |\n| CI environment (`$CI` set) | No |\n| `KODELYTH_NO_MENU=1 kodelythecc` | No |\n\n### Uninstall\n\nThe menu's **Uninstall ECC completely** row runs an interactive full cleanup: removes the 777 ECC-installed files from `~/.claude/`, unwires RTK from your AI tool, removes codebase-memory-mcp agent configs, removes ECC's MCP entry from Claude Code + Claude Desktop, and deletes `~/.kodelythecc/` (memory + ledgers). Prompts confirm before anything is deleted; a dry-run mode previews what would be removed without touching anything.\n\nYou can also run it non-interactively:\n\n```bash\nkodelythecc uninstall --dry-run # preview what would be removed\nkodelythecc uninstall --yes # full cleanup\nkodelythecc uninstall --yes --keep-memory # remove files but keep ~/.kodelythecc/\nnpm uninstall -g kodelyth-ecc # finally remove the npm package itself\n```\n\n### Update check\n\nThe menu polls `https://registry.npmjs.org/kodelyth-ecc/latest` on open. Cached 24h in `~/.kodelythecc/update-check.json`. When a newer version exists, an extra menu row appears at the top:\n\n```\n ▸ Update to v2.3.1 [NEW] npm i -g kodelyth-ecc\n```\n\n### Background daemon\n\nSelecting \"Run in background\" forks the dashboard as a detached process:\n\n```\n✓ Dashboard daemon started (pid 12345)\n URL: http://127.0.0.1:5747\n Log: /Users/you/.kodelythecc/dashboard-daemon.log\n Pid: /Users/you/.kodelythecc/dashboard-daemon.pid\n Stop: kill $(cat ~/.kodelythecc/dashboard-daemon.pid)\n```\n\nSurvives shell exit. Real daemon.\n\n---\n\n## Kodelyth Memory — Local Self-Learning\n\n
\n\n
\n\nThe first time you solve a hard problem, ECC remembers what worked. The next time you hit something similar, your AI surfaces the past solution before you ask.\n\n```\nSession 1 (March):\n You: \"Stripe webhook signatures are failing in production\"\n AI: [helps you debug, you discover raw body parser is required]\n You: \"Perfect, that worked, thanks\"\n → ECC's Stop hook queues the lesson for review.\n → You confirm with /memory review-pending → stored locally.\n\nSession 2 (August, new project):\n You: \"I need to add Stripe webhooks to checkout\"\n AI: I checked your memory — you solved this exact problem in March.\n Raw body parser before signature validation, test with stripe-cli\n not curl. Want me to apply the same pattern?\n You: \"Yes, do it\"\n → 30 minutes of past debugging saved in 3 seconds.\n```\n\n### How it works\n\n| Layer | Mechanism |\n|---|---|\n| **Capture** | A Stop hook scans your session JSONL, extracts (problem, approach, gotchas, tags). Queues for your review — never auto-stores. |\n| **Storage** | `~/.kodelyth/memory/memories.jsonl` — append-only log on your disk only. Override with `KODELYTH_MEMORY_DIR`. |\n| **Retrieval** | BM25 keyword + tag search. Pure JS, sub-millisecond, no embeddings, no network. |\n| **Session-start injection** | A SessionStart hook builds a cache-friendly context block: stable prefix (your patterns + recent project memories) → variable suffix (relevant to current task). |\n| **Auto chat detection** | A `UserPromptSubmit` hook watches every message you type, runs BM25 search on it, and injects relevant memories **before the AI responds**. Per-session repeat suppression. |\n| **Cost win** | The stable prefix sits in the prompt cache. Anthropic charges 10% on cached tokens (5-min TTL); OpenAI auto-caches prefixes ≥1024 tokens. Long sessions become dramatically cheaper. |\n\n### Privacy\n\nEvery byte stays on your machine. Verify any time with `ls -la ~/.kodelyth/memory/`. Sync across machines is opt-in (Dropbox/iCloud/git on that folder).\n\n### Slash command\n\n```\n/memory # Stats and recent memories\n/memory recall # BM25 search\n/memory remember \"\" # Capture (interactive — confirms before storing)\n/memory review-pending # Review Stop-hook candidates\n/memory forget <id> # Soft-delete a memory\n/memory inject [--query <text>] # Print what your AI sees about you\n```\n\n### Honest limits\n\n- It is **not** model fine-tuning. The LLM never changes. We give it smarter context.\n- Cache savings apply to Anthropic and OpenAI. Other models (Gemini, Llama, Mistral) get the recall quality without the discount.\n- Cloud-AI platforms (Windsurf, Antigravity, partial Cursor) store sessions server-side. Auto-extract from past sessions doesn't work there. Manual `/memory remember` still does.\n\nSee `skills/kodelyth-memory/SKILL.md` for the full design + CLI reference.\n\n---\n\n## Compound Learning System — Self-Improvement Loop\n\n<div align=\"center\">\n<img src=\"social/section-learning.svg\" alt=\"Compound learning — every correction stacks, Claude matches your style\" width=\"900\"/>\n</div>\n\nEvery correction you give Claude gets encoded into your project permanently. The toolkit gets smarter every session without any effort from you.\n\n```\nSession 1: You type \"use pnpm not npm\"\n → Session ends\n → capture-correction.js scans the JSONL\n → Writes to tasks/lessons.md: \"- use pnpm not npm\"\n\nSession 2: read-lessons.js fires at session start\n → Injects: \"PROJECT LESSONS — HARD RULES: - use pnpm not npm\"\n → Claude uses pnpm without being told\n\nMonth 1: 10+ corrections stacked\n → Claude knows your naming style, preferred patterns, tech opinions\n → Zero ramp-up time on any new task\n\nMonth 3: You try another tool\n → It uses npm. It uses the wrong pattern. It asks basic questions.\n → You come back.\n```\n\n### Three-layer compound memory architecture\n\n| Layer | File | Scope | How it works |\n|---|---|---|---|\n| **Project Lessons** | `tasks/lessons.md` | Per-project | Hard rules from your corrections. Injected at session start as mandatory context. |\n| **Global Memory** | `~/.kodelyth/memory/` | Cross-project | BM25 fuzzy recall of past solutions. Auto-fires on every prompt you type. |\n| **Intent Routing** | 70 agents | Always-on | Routes your message to the right specialist from the first word. No names needed. |\n\n### How it works under the hood\n\n**`capture-correction.js`** — Stop hook, runs async at session end:\n- Scans session JSONL for 12 correction signal patterns (`\"no don't\"`, `\"use X instead\"`, `\"stop doing Y\"`, `\"we always\"`, `\"wrong approach\"`, etc.)\n- Extracts them as plain-language rules\n- Appends them to `tasks/lessons.md` with date grouping\n\n**`read-lessons.js`** — SessionStart hook, fires first before any other hook:\n- Reads `tasks/lessons.md` and formats rules as `PROJECT LESSONS — HARD RULES` block\n- Detects project DNA automatically: Node.js + framework (Next.js, React, NestJS, etc.), Go, Rust, Python, Java/Gradle, package manager (pnpm/bun/yarn/npm), test runner\n- Surfaces open `tasks/todo.md` items into session context\n\n### `tasks/lessons.md` — your project's rulebook\n\nEdit it freely. Add rules manually. Remove rules that no longer apply. It's a plain markdown file at `tasks/lessons.md` in your project root.\n\n```markdown\n# Claude Lessons\n\nProject: **my-app**\n\n## 2026-05-06\n\n- use pnpm not npm\n- never add try/catch without logging the error first\n- we use Zod for validation, not Yup\n- component files go in src/components, not src/app\n```\n\n---\n\n## What's Inside\n\n| Component | Count | Description |\n|---|---|---|\n| Agents | **70** | Specialist subagents — reviewers, planners, debuggers, architects, devil-mode adversarial crew, incident-commander, load-tester, memory, image-architect |\n| Skills | **194** | Domain knowledge — patterns, testing, security, DevOps, intent routing, memory |\n| Commands | **97** | Slash command workflows (`/tdd`, `/plan`, `/memory`, `/devil-mode`, `/doctor`, `/update`, etc.) |\n| Parallel commands | **8** | `/devil-mode`, `/team-review`, `/security-audit`, `/debug-blitz`, `/refactor-sprint`, `/pre-release`, `/onboard`, `/project-launch` |\n| Hooks | **22+** | Quality gates, secret scanning, branch checks, memory inject + capture + correction + project DNA |\n| Rules | **14** | Always-on coding standards + intent routing + memory protocol + self-improvement workflow |\n| Memory | **local** | BM25-indexed personal memory at `~/.kodelyth/memory/` (zero deps) |\n\n---\n\n## Agent Arsenal\n\n<div align=\"center\">\n<img src=\"social/section-agents.svg\" alt=\"70 specialist agents — one for every situation\" width=\"900\"/>\n</div>\n\n<div align=\"center\">\n<img src=\"social/readme-agents.svg\" alt=\"70 specialist agents — agent grid preview\" width=\"900\"/>\n</div>\n\n### Kodelyth Exclusives — The 16 Agents That Make ECC\n\n| Agent | One-line job |\n|---|---|\n| `kodelyth-advisor` | Master guide — picks the right tool when you don't know where to start |\n| `kodelyth-memory` | Curates your local memory — recalls past solutions, captures new ones |\n| `pair-programmer` | The engineer who sits next to you **before** you write the code |\n| `debug-detective` | Never guesses — traces every bug to root cause |\n| `silent-failure-hunter` | Finds bugs that don't throw errors |\n| `incident-commander` | Runs production incidents — triage, contain, postmortem. P0s only. |\n| `load-tester` | Load and stress testing — k6, Locust, Artillery, capacity planning |\n| `ux-reviewer` | Reviews UX behavior + WCAG 2.1 AA accessibility (never touches design) |\n| `api-guardian` | Detects breaking API changes before they ship |\n| `migration-guide` | Framework / language version upgrades, phase by phase |\n| `dependency-doctor` | npm/pip/cargo/maven hell — CVE triage, lockfile diagnosis |\n| `git-rescue` | Recovers from broken git states without destroying history |\n| `release-captain` | Owns the release ritual — semver, tagging, publishing, rollback |\n| `env-debugger` | \"Works on my machine\" hunter — env, config, secrets, layers |\n| `flake-hunter` | Stabilizes flaky tests — never adds blind retries |\n| `image-architect` | AI image generation — Gemini/DALL-E/fal.ai/SVG, platform-aware |\n\n### Agent Categories\n\n| Category | Agents |\n|---|---|\n| **Guidance** | `kodelyth-advisor`, `pair-programmer`, `planner`, `architect`, `code-architect`, `chief-of-staff`, `migration-guide` |\n| **Code Review** | `code-reviewer`, `typescript-reviewer`, `python-reviewer`, `go-reviewer`, `rust-reviewer`, `java-reviewer`, `kotlin-reviewer`, `cpp-reviewer`, `csharp-reviewer`, `flutter-reviewer`, `database-reviewer`, `healthcare-reviewer` |\n| **Build Fixers** | `build-error-resolver`, `go-build-resolver`, `rust-build-resolver`, `java-build-resolver`, `kotlin-build-resolver`, `cpp-build-resolver`, `dart-build-resolver`, `pytorch-build-resolver`, `dependency-doctor`, `env-debugger` |\n| **Debugging** | `debug-detective`, `silent-failure-hunter`, `flake-hunter` |\n| **Incident & Load** | `incident-commander`, `load-tester` |\n| **Security & API** | `security-reviewer`, `api-guardian` |\n| **Performance** | `performance-optimizer` |\n| **Quality** | `refactor-cleaner`, `code-simplifier`, `type-design-analyzer` |\n| **Testing** | `tdd-guide`, `e2e-runner`, `pr-test-analyzer`, `flake-hunter` |\n| **Documentation** | `doc-updater`, `docs-lookup`, `comment-analyzer` |\n| **Release & Ops** | `release-captain`, `git-rescue` |\n| **Open Source** | `opensource-forker`, `opensource-packager`, `opensource-sanitizer` |\n| **Specialized** | `seo-specialist`, `ux-reviewer`, `code-explorer` |\n\n---\n\n## Hooks — Always Running in Background\n\n<div align=\"center\">\n<img src=\"social/section-hooks.svg\" alt=\"22+ automated quality hooks — PreToolUse, PostToolUse, Stop\" width=\"900\"/>\n</div>\n\nOnce installed (Claude Code target), these hooks run automatically with zero configuration:\n\n| Hook | What it does |\n|---|---|\n| Session start — lessons | Reads `tasks/lessons.md` and injects your hard rules + project DNA as context |\n| Session start — memory | Loads relevant past solutions from global BM25 memory |\n| Auto chat recall | Watches every prompt, injects relevant memories before AI responds |\n| Correction capture | Detects when you correct Claude, encodes the rule to `tasks/lessons.md` |\n| Pre-commit | Catches `console.log`, secrets, bad commit messages |\n| Quality gate | Runs type checks and formatting after edits |\n| Git push reminder | Prompts review before pushing |\n| Config protection | Blocks weakening of linter/formatter configs |\n| Desktop notify | macOS notification when a long task finishes |\n| MCP health check | Validates MCP servers before calling them |\n| Test reminder | Prompts to write tests when code is edited without tests |\n| Branch name check | Blocks git branches that don't match naming convention |\n\n---\n\n## Usage After Install\n\n### Start here (new users)\n\n```\n/kodelyth-quickstart\n```\n\n### Core workflows — explicit invocation\n\n```bash\nuse kodelyth-advisor # Not sure what to do? Start here\nuse pair-programmer # Think through approach before writing code\nuse planner # Plan a feature before writing code\nuse code-reviewer # Review after writing code\nuse debug-detective # Trace a bug to its root cause\nuse api-guardian # Check API changes for breaking contracts\nuse security-reviewer # Review security-sensitive code\nuse ux-reviewer # Review frontend UX and accessibility\nuse migration-guide # Upgrading a framework or language version\nuse tdd-guide # Write tests the right way\nuse performance-optimizer # Diagnose and fix slowness\nuse dependency-doctor # npm/pip/cargo dep hell\nuse git-rescue # Broken git state, lost commits, bad rebase\nuse release-captain # Cut a clean release with rollback plan\nuse env-debugger # \"Works on my machine\" — env/config/secrets\nuse flake-hunter # Stabilize flaky tests\n```\n\n### Or just describe your problem\n\nThe intent router will route you to the right one. The AI announces who's taking over, helps you, and tells you the explicit name for next time.\n\n### Load language patterns (skills)\n\n```bash\n/typescript-patterns # TypeScript + React + Next.js\n/python-patterns # Python best practices\n/golang-patterns # Go best practices\n/postgres-patterns # Database patterns\n/docker-patterns # Container patterns\n/coding-standards # Universal baseline\n```\n\n### Advanced skills\n\n```bash\n/git-mastery # Trunk-based dev, rebase, bisect, monorepos\n/observability # Structured logging, metrics, OpenTelemetry, SLOs\n/smart-debug # Systematic debugging framework\n/intent-routing # How the auto-routing system works\n/agent-handoff # Standard multi-agent handoff chains\n```\n\n---\n\n## What Gets Installed\n\n### Claude Code — `~/.claude/`\n\n| Source | Destination | What it does |\n|---|---|---|\n| `agents/` | `~/.claude/agents/` | All 70 subagents available globally |\n| `skills/` | `~/.claude/skills/` | All 196 skills loadable via commands |\n| `hooks/hooks.json` | `~/.claude/hooks/` | Automated quality gates |\n| `rules/` | `~/.claude/rules/` | Always-on standards + intent routing |\n| `commands/` | `~/.claude/commands/` | Slash commands (`/tdd`, `/plan`, etc.) |\n\n### Windsurf — `.windsurf/` (project) or `~/.codeium/windsurf/` (global)\n\n| Source | Destination | Notes |\n|---|---|---|\n| `agents/` | `.windsurf/agents/` | Call with `use <agent-name>` in Cascade |\n| `skills/` | `.windsurf/skills/` | Domain knowledge, loadable via chat |\n| `rules/` | `.windsurf/rules/` | Flattened rule files |\n| `rules/common/` | `.windsurfrules` | Concatenated — Windsurf reads this automatically every session |\n\n### Cursor — `.cursor/` in project root\n\n| Source | Destination |\n|---|---|\n| `rules/` | `.cursor/rules/` |\n| `skills/` | `.cursor/skills/` |\n\n### Codex CLI — `~/.codex/`\n\n| Source | Destination |\n|---|---|\n| `agents/` | `~/.codex/agents/` |\n| `skills/` | `~/.codex/skills/` |\n| `commands/` | `~/.codex/commands/` |\n| `rules/` | `~/.codex/rules/` |\n\n### Antigravity — `.agent/` in project root\n\n| Source | Destination |\n|---|---|\n| `agents/` | `.agent/skills/` |\n| `commands/` | `.agent/workflows/` |\n| `rules/` | `.agent/rules/` |\n\n### OpenCode — `.opencode/` in project root\n\n| Source | Destination |\n|---|---|\n| `rules/` | `.opencode/rules/` |\n\n---\n\n## Install Profiles — Language Bundles\n\nWorks with any target:\n\n```bash\nnpx kodelyth-ecc --profile nextjs # Claude Code\nnpx kodelyth-ecc --target windsurf-project --profile nextjs # Windsurf\nnpx kodelyth-ecc --target antigravity --profile python-api # Antigravity\n```\n\nAvailable profiles:\n\n| Profile | Includes |\n|---|---|\n| `nextjs` | TypeScript + React + Next.js rules |\n| `python-api` | Python + Django/FastAPI rules |\n| `fullstack` | TypeScript + Python + Go rules |\n| `mobile` | Kotlin + Swift rules |\n| `backend` | Go + Python + Java rules |\n\nOr specify languages directly:\n\n```bash\nnpx kodelyth-ecc typescript python golang rust java kotlin php swift cpp dart ruby elixir\n```\n\n---\n\n## Multi-Platform Support\n\n| Platform | Supported | Install target | What gets installed |\n|---|---|---|---|\n| Claude Code | Full | `claude-home` (default) | Agents, skills, commands, hooks, rules |\n| Windsurf | Full | `windsurf-project` / `windsurf-home` | Agents, skills, rules, `.windsurfrules` |\n| Cursor | Full | `cursor-project` | Rules, skills |\n| Codex CLI | Full | `codex-home` | Agents, skills, commands, rules |\n| Google Antigravity | Full | `antigravity` | Agents → skills, commands → workflows, rules |\n| OpenCode | Rules only | `opencode` | Rules (agents + skills not yet supported by OpenCode) |\n\n**OS support:** macOS, Linux (`install.sh`), Windows (`install.ps1`), or any OS with Node.js 18+ (`npx`).\n\n---\n\n## Privacy & Philosophy\n\nECC is **100% local files**. No telemetry, no cloud, no account, no tracking. Everything is markdown your AI reads on every session.\n\n- **No phoning home** — install scripts copy files and exit. MCP server, dashboard, swarm, and replay all stay on localhost.\n- **You own everything** — fork the repo, edit any agent, write your own.\n- **Verifiable** — `npx kodelyth-ecc verify` checks your install against the sha256 manifest.\n\n---\n\n## Contributing\n\nWe welcome new agents, skills, hooks, and intent routing patterns that meet the **Kodelyth Standard** — specialist personas, model-agnostic, production-grade examples.\n\nWhen adding a new agent, also update `rules/common/agent-intent-routing.md` with your trigger patterns so the intent router knows when to call you.\n\nSee [CONTRIBUTING.md](CONTRIBUTING.md) for templates, checklists, and the full Kodelyth Standard.\n\n---\n\n## Changelog\n\nSee [CHANGELOG.md](CHANGELOG.md). v1.8.0 highlights:\n\n- **Visual system overhaul** — `github-social-preview.svg` and `og-image.svg` fully redesigned with two-panel layout: brand left, 2×2 stat cards right, dot grid texture, amber accent\n- **SVG overflow fixes** — KODELYTH wordmark, bottom bar text, and stats numerals all corrected to stay within canvas bounds across all 31 social assets\n- **Test count update** — 373 tests across 25 test files (up from 348)\n- **4K PNG exports** — all 31 social assets regenerated at 4× scale; fullhd exports removed\n\nv1.7.3 highlights:\n\n- **Social hype pack** — 5 X/Twitter SVG cards + 5 thread scripts, full SVG version sync, GitHub issue templates, repo polish\n\nv1.7.0 highlights:\n\n- **MCP Server** — 16 tools, 6 prompts, 377 resources; stdio JSON-RPC; zero telemetry; works with Claude Desktop, LangGraph, AutoGen, CrewAI, OpenAI Agents SDK\n- **Dashboard hardening** — DNS rebinding protection, shell injection fix (`execFileSync`), CSP on all responses, SSE client cap (max 10), TOCTOU fixes\n- **Devil Mode** — 8 adversarial agents, `--pre-public` / `--pre-launch` / `--all` flags, real bash-grep patterns and remediation playbooks\n- **336 tests** passing across 25 test files (now 373 at v1.8.0)\n\nv1.5.8 highlights:\n\n- **zsh inline comment fix** — `npx kodelyth-ecc --target windsurf-home # comment` no longer crashes.\n- **Dynamic install counts** — agent/skill/command counts computed from filesystem at install time.\n- **Ruby + Elixir language rules** — coding-style, patterns, testing, security, and hooks.\n- **`/doctor` command** — health check on your ECC install without leaving your IDE.\n- **`/update` command** — upgrade to the latest version automatically.\n- **5 new parallel commands** (`/security-audit`, `/pre-release`, `/debug-blitz`, `/refactor-sprint`, `/onboard`).\n\nv1.5.2 highlights:\n\n- **Added:** `image-architect` agent — platform-aware AI image generation (Gemini Imagen 3, DALL-E 3, fal.ai, SVG)\n- **Added:** `/project-launch` — 5 founding-team agents fire in parallel (45 min → 10 min)\n- **Added:** `/team-review` — 4 audit agents fire in parallel (60 min → 15 min)\n- **Added:** `/lessons` command — cross-platform lesson loading\n\nv1.5.1 highlights:\n\n- **Added:** `capture-correction.js` Stop hook — auto-encodes corrections to `tasks/lessons.md`\n- **Added:** `read-lessons.js` SessionStart hook — injects lessons + project DNA before first message\n- **Added:** `rules/common/self-improvement-workflow.md` — 6-pattern self-improvement loop\n\nv1.5.0 highlights:\n\n- **Added:** `incident-commander` — production incident response\n- **Added:** `load-tester` — load/stress testing with k6, Locust, Artillery\n\nv1.4.0 highlights:\n\n- **Added:** `kodelyth-memory` — local self-learning memory (BM25, zero deps, cache-friendly)\n- **Added:** Auto chat detection hook, SessionStart injection, Stop hook queue\n\n---\n\n## Author\n\n<div align=\"center\">\n<img src=\"social/section-author.svg\" alt=\"Shofiqul Islam — Creator of Kodelyth ECC\" width=\"900\"/>\n</div>\n\n**Kodelyth ECC** is designed and maintained by **Shofiqul Islam** — full-stack engineer and AI tooling builder.\n\n| Platform | Link |\n|---|---|\n| GitHub | [@sifxprime](https://github.com/sifxprime) |\n| X / Twitter | [@sifxprime](https://x.com/sifxprime) |\n| Facebook | [facebook.com/sifxprime](https://facebook.com/sifxprime) |\n| Instagram | [@sifxprime](https://instagram.com/sifxprime) |\n| npm | [npmjs.com/package/kodelyth-ecc](https://www.npmjs.com/package/kodelyth-ecc) |\n\n---\n\n## Support\n\nReal people, and we answer.\n\n| Channel | Reach us |\n|---|---|\n| Email | **[ecc@kodelyth.com](mailto:ecc@kodelyth.com)** |\n| WhatsApp | **[+880 1312 365939](https://wa.me/8801312365939)** |\n| Bug reports | [GitHub Issues](https://github.com/sifxprime/kodelyth-ecc/issues) |\n| Docs | [ecc.kodelyth.com/docs](https://ecc.kodelyth.com/docs) |\n\n**Found a security issue?** Email [ecc@kodelyth.com](mailto:ecc@kodelyth.com)\ndirectly rather than opening a public issue, and give us a reasonable window to\nship a fix before disclosing it.\n\nWhen reporting a bug, the three things that get it fixed fastest are the exact\ncommand you ran, the full error, and your `kodelythecc doctor` output.\n\n---\n\n<div align=\"center\">\n\n[](https://www.npmjs.com/package/kodelyth-ecc)\n[](https://github.com/sifxprime/kodelyth-ecc)\n[](LICENSE)\n[](https://nodejs.org)\n\nBuilt with craft. Zero telemetry. All yours.\n\n</div>\n","readmeFilename":"README.md"}
![\"Kodelyth](\"brand/ecc-lockup-dark.svg\")
\n![\"Kodelyth](\"social/readme-hero.svg\")
\n\n[](https://www.npmjs.com/package/kodelyth-ecc)\n[](https://www.npmjs.com/package/kodelyth-ecc)\n[](https://github.com/sifxprime/kodelyth-ecc/stargazers)\n[](LICENSE)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n![\"Install](\"social/section-install.svg\")
\n![\"MCP](\"social/section-mcp.svg\")
\n![\"Intent](\"social/section-routing.svg\")
\n![\"Parallel](\"social/section-parallel.svg\")
\n![\"Devil](\"social/section-devil.svg\")
\n![\"Dashboard](\"social/section-dashboard.svg\")
\n![\"RTK](\"social/section-rtk.svg\")
![\"Terse](\"social/section-terse.svg\")
![\"Codebase](\"social/section-codebase.svg\")
![\"Interactive](\"social/section-cli.svg\")
![\"Kodelyth](\"social/section-memory.svg\")
\n