kodelyth-ecc 1.2.2 → 1.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/AGENTS.md +101 -181
- package/CHANGELOG.md +67 -0
- package/CLAUDE.md +72 -63
- package/KODELYTH.md +79 -44
- package/README.md +244 -192
- package/VERSION +1 -1
- package/agents/dependency-doctor.md +120 -0
- package/agents/env-debugger.md +154 -0
- package/agents/flake-hunter.md +142 -0
- package/agents/git-rescue.md +133 -0
- package/agents/kodelyth-memory.md +87 -0
- package/agents/release-captain.md +190 -0
- package/bin/kodelyth-ecc.js +18 -12
- package/commands/memory.md +62 -0
- package/hooks/hooks.json +26 -0
- package/hooks/memory/capture-stop.js +88 -0
- package/hooks/memory/inject-start.js +60 -0
- package/install.ps1 +28 -9
- package/install.sh +11 -97
- package/package.json +4 -2
- package/rules/common/agent-intent-routing.md +337 -0
- package/rules/common/memory-protocol.md +56 -0
- package/scripts/memory/cli.js +200 -0
- package/scripts/memory/extract.js +176 -0
- package/scripts/memory/inject.js +145 -0
- package/scripts/memory/store.js +300 -0
- package/skills/agent-handoff/SKILL.md +184 -0
- package/skills/intent-routing/SKILL.md +134 -0
- package/skills/kodelyth-memory/SKILL.md +136 -0
- package/tests/memory/store.test.js +121 -0
- package/dashboard/lib/agent-tracker.js +0 -366
- package/dashboard/lib/aggregator.js +0 -119
- package/dashboard/lib/cost-calculator.js +0 -50
- package/dashboard/lib/platform-detector.js +0 -89
- package/dashboard/lib/readers/antigravity-reader.js +0 -113
- package/dashboard/lib/readers/claude-reader.js +0 -135
- package/dashboard/lib/readers/codex-reader.js +0 -192
- package/dashboard/lib/readers/cursor-reader.js +0 -135
- package/dashboard/lib/readers/opencode-reader.js +0 -201
- package/dashboard/lib/readers/windsurf-reader.js +0 -146
- package/dashboard/package.json +0 -24
- package/dashboard/public/index.html +0 -1221
- package/dashboard/server.js +0 -119
- package/scripts/agent-tracker-hook.js +0 -81
- package/social/readme-lens.svg +0 -140
- package/social/readme-savings.svg +0 -56
package/README.md
CHANGED
|
@@ -4,57 +4,70 @@
|
|
|
4
4
|
|
|
5
5
|
[](LICENSE)
|
|
6
6
|

|
|
7
|
+

|
|
7
8
|

|
|
8
|
-

|
|
10
|
+

|
|
11
|
+

|
|
12
|
+

|
|
13
|
+

|
|
14
|
+

|
|
11
15
|
|
|
12
16
|
</div>
|
|
13
17
|
|
|
14
|
-
**Kodelyth ECC** is a
|
|
18
|
+
**Kodelyth ECC** is a production-grade AI coding toolkit — **59 specialist agents, 188 skills, 80 commands**, a god-tier **intent-routing system**, and **local self-learning memory** that gets smarter at helping *you* every session.
|
|
15
19
|
|
|
16
|
-
Works with **Claude Code**, **
|
|
20
|
+
Works with **Claude Code**, **Windsurf**, **Cursor**, **Codex CLI**, **Google Antigravity**, and **OpenCode**.
|
|
21
|
+
|
|
22
|
+
> No dashboard. No telemetry. No cloud. Just rules, agents, skills, and your own private memory store — all on your disk.
|
|
17
23
|
|
|
18
24
|
---
|
|
19
25
|
|
|
20
|
-
##
|
|
26
|
+
## What Makes ECC Different
|
|
21
27
|
|
|
22
|
-
|
|
28
|
+
Most "AI agent kits" are folders of markdown files you have to remember the names of. ECC has **intent routing** baked into the rules layer — you describe your problem in plain words, the AI announces which specialist is taking over, and you get senior-grade help without memorizing a directory.
|
|
23
29
|
|
|
24
|
-
|
|
30
|
+
```
|
|
31
|
+
You: "I've been staring at this NullPointerException for two hours,
|
|
32
|
+
I'm losing my mind."
|
|
25
33
|
|
|
26
|
-
|
|
27
|
-
|
|
34
|
+
AI: → Routing to debug-detective (your error + frustration matches the bug-tracking signal)
|
|
35
|
+
|
|
36
|
+
That kind of bug is exhausting — let's trace it properly so we
|
|
37
|
+
fix the root cause, not the symptom.
|
|
38
|
+
|
|
39
|
+
First, can you share the full stack trace and...
|
|
28
40
|
```
|
|
29
41
|
|
|
30
|
-
|
|
42
|
+
You never typed `use debug-detective`. You didn't have to.
|
|
43
|
+
|
|
44
|
+
---
|
|
45
|
+
|
|
46
|
+
## Install
|
|
47
|
+
|
|
48
|
+
### Option 1 — npx from npm (recommended, any platform)
|
|
31
49
|
|
|
32
50
|
```bash
|
|
33
|
-
npx kodelyth-ecc
|
|
51
|
+
npx kodelyth-ecc # Claude Code (default)
|
|
52
|
+
npx kodelyth-ecc --target windsurf-project # Windsurf (per-project)
|
|
34
53
|
npx kodelyth-ecc --target windsurf-home # Windsurf (global)
|
|
35
|
-
npx kodelyth-ecc --target antigravity # Google Antigravity
|
|
36
54
|
npx kodelyth-ecc --target cursor-project # Cursor IDE
|
|
37
55
|
npx kodelyth-ecc --target codex-home # Codex CLI
|
|
56
|
+
npx kodelyth-ecc --target antigravity # Google Antigravity
|
|
38
57
|
npx kodelyth-ecc --target opencode # OpenCode
|
|
39
58
|
```
|
|
40
59
|
|
|
41
|
-
|
|
42
|
-
|
|
43
|
-
---
|
|
60
|
+
Node.js 18+ required. [Download Node](https://nodejs.org) if you don't have it.
|
|
44
61
|
|
|
45
62
|
### Option 2 — npx from GitHub (always latest commit)
|
|
46
63
|
|
|
47
|
-
Use this if you want the very latest unreleased version directly from the main branch.
|
|
48
|
-
|
|
49
64
|
```bash
|
|
50
65
|
npx github:sifxprime/kodelyth-ecc
|
|
51
66
|
```
|
|
52
67
|
|
|
53
|
-
Same `--target` flags
|
|
54
|
-
|
|
55
|
-
---
|
|
68
|
+
Same `--target` flags work.
|
|
56
69
|
|
|
57
|
-
### Option 3 — curl (macOS / Linux)
|
|
70
|
+
### Option 3 — curl (macOS / Linux only)
|
|
58
71
|
|
|
59
72
|
```bash
|
|
60
73
|
curl -fsSL https://raw.githubusercontent.com/sifxprime/kodelyth-ecc/main/install.sh | bash
|
|
@@ -66,198 +79,190 @@ With a target:
|
|
|
66
79
|
curl -fsSL https://raw.githubusercontent.com/sifxprime/kodelyth-ecc/main/install.sh | bash -s -- --target windsurf-project
|
|
67
80
|
```
|
|
68
81
|
|
|
69
|
-
|
|
70
|
-
|
|
71
|
-
### Option 4 — clone and run (all platforms)
|
|
82
|
+
### Option 4 — Clone and run
|
|
72
83
|
|
|
73
84
|
```bash
|
|
74
85
|
git clone https://github.com/sifxprime/kodelyth-ecc.git
|
|
75
86
|
cd kodelyth-ecc
|
|
76
87
|
|
|
77
88
|
# macOS / Linux
|
|
78
|
-
./install.sh
|
|
79
|
-
./install.sh --target windsurf-project
|
|
80
|
-
./install.sh --target antigravity # Google Antigravity
|
|
89
|
+
./install.sh # Claude Code (default)
|
|
90
|
+
./install.sh --target windsurf-project # Windsurf
|
|
81
91
|
|
|
82
92
|
# Windows (PowerShell)
|
|
83
93
|
.\install.ps1
|
|
84
94
|
.\install.ps1 -Target windsurf-project
|
|
85
|
-
.\install.ps1 -Target codex-home
|
|
86
95
|
```
|
|
87
96
|
|
|
88
97
|
---
|
|
89
98
|
|
|
90
|
-
|
|
91
|
-
|
|
92
|
-
---
|
|
93
|
-
|
|
94
|
-
## Kodelyth Lens — AI Usage Dashboard
|
|
99
|
+
## How Intent Routing Works
|
|
95
100
|
|
|
96
|
-
|
|
101
|
+
The 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.
|
|
97
102
|
|
|
98
|
-
|
|
103
|
+
### Two activation paths
|
|
99
104
|
|
|
100
|
-
|
|
105
|
+
**1. Explicit** — type it directly:
|
|
106
|
+
```
|
|
107
|
+
use debug-detective
|
|
108
|
+
@code-reviewer
|
|
109
|
+
invoke security-reviewer
|
|
110
|
+
```
|
|
101
111
|
|
|
102
|
-
|
|
112
|
+
**2. Implicit** — just describe your problem; the AI routes you:
|
|
103
113
|
|
|
104
|
-
|
|
|
114
|
+
| What you write | Auto-routed to |
|
|
105
115
|
|---|---|
|
|
106
|
-
|
|
|
107
|
-
|
|
|
108
|
-
|
|
|
109
|
-
|
|
|
110
|
-
|
|
|
111
|
-
|
|
|
116
|
+
| "I'm stuck, no idea where to start" | `kodelyth-advisor` |
|
|
117
|
+
| "I've been debugging this for hours" | `debug-detective` |
|
|
118
|
+
| "Should I use Context or Zustand?" | `pair-programmer` |
|
|
119
|
+
| "Build failed on Vercel" | `build-error-resolver` |
|
|
120
|
+
| "Is this JWT signing secure?" | `security-reviewer` |
|
|
121
|
+
| "Why is this so slow?" | `performance-optimizer` |
|
|
122
|
+
| "Plan the v2 migration" | `planner` + `migration-guide` |
|
|
123
|
+
| "Tests pass locally but fail on CI" | `flake-hunter` + `env-debugger` |
|
|
124
|
+
| "I lost my commits after `reset --hard`" | `git-rescue` |
|
|
125
|
+
| "npm install is failing" | `dependency-doctor` |
|
|
126
|
+
| "Cut a 1.4 release" | `release-captain` |
|
|
127
|
+
| "Add accessibility to this form" | `ux-reviewer` |
|
|
128
|
+
| "Open-source this project" | `opensource-forker` (chain) |
|
|
129
|
+
|
|
130
|
+
The AI **always announces** which agent is taking over (`→ Routing to <agent>`) and **always teaches** you the explicit form for next time (`Tip: type "use <agent>"`). No silent personality changes.
|
|
131
|
+
|
|
132
|
+
### Multi-agent chains
|
|
133
|
+
|
|
134
|
+
Real problems span multiple specialties. ECC ships standard handoff chains:
|
|
112
135
|
|
|
113
|
-
### Supported platforms
|
|
114
|
-
|
|
115
|
-
| Platform | Data source | Sessions |
|
|
116
|
-
|---|---|---|
|
|
117
|
-
| Claude Code | `~/.claude/projects/**/*.jsonl` | Full local logs |
|
|
118
|
-
| Codex CLI | `~/.codex/state_5.sqlite` | SQLite threads table |
|
|
119
|
-
| Windsurf | `~/Library/Application Support/Windsurf/` | Cloud AI (server-side) |
|
|
120
|
-
| Antigravity | `~/Library/Application Support/Antigravity/` | Cloud AI (server-side) |
|
|
121
|
-
| Cursor | `~/.cursor/` | Local workspace data |
|
|
122
|
-
| OpenCode | `~/.opencode/` | Local session files |
|
|
123
|
-
|
|
124
|
-
Only installed platforms appear in the dashboard — nothing shown for platforms you don't have.
|
|
125
|
-
|
|
126
|
-
### Manual start
|
|
127
|
-
|
|
128
|
-
```bash
|
|
129
|
-
node ~/.kodelyth/lens/server.js
|
|
130
|
-
# → http://localhost:3456
|
|
131
136
|
```
|
|
132
|
-
|
|
133
|
-
|
|
134
|
-
|
|
135
|
-
```bash
|
|
136
|
-
CLAUDE_DIR=~/.claude # Custom Claude Code data dir
|
|
137
|
-
CURSOR_DIR=... # Custom Cursor data dir
|
|
138
|
-
WINDSURF_DIR=... # Custom Windsurf data dir
|
|
139
|
-
CODEX_DIR=~/.codex # Custom Codex data dir
|
|
140
|
-
OPENCODE_DIR=~/.opencode # Custom OpenCode data dir
|
|
141
|
-
ANTIGRAVITY_DIR=... # Custom Antigravity data dir
|
|
142
|
-
PORT=3456 # Dashboard port (default: 3456)
|
|
137
|
+
pair-programmer → tdd-guide → code-reviewer → security-reviewer
|
|
138
|
+
(approach) (write tests) (review impl) (auth, validation)
|
|
143
139
|
```
|
|
144
140
|
|
|
145
|
-
|
|
141
|
+
```
|
|
142
|
+
debug-detective → tdd-guide → refactor-cleaner
|
|
143
|
+
(root cause) (regression test) (cleanup)
|
|
144
|
+
```
|
|
146
145
|
|
|
147
|
-
|
|
146
|
+
```
|
|
147
|
+
opensource-forker → opensource-sanitizer → opensource-packager → release-captain
|
|
148
|
+
(clean fork) (strip secrets) (README, license) (cut v0.1.0)
|
|
149
|
+
```
|
|
148
150
|
|
|
149
|
-
|
|
150
|
-
|---|---|---|
|
|
151
|
-
| Agents | **53** | Specialist subagents — reviewers, planners, debuggers, architects |
|
|
152
|
-
| Skills | **185** | Domain knowledge — patterns, testing, security, DevOps |
|
|
153
|
-
| Commands | **79** | Slash command workflows (`/tdd`, `/plan`, `/code-review`, etc.) |
|
|
154
|
-
| Hooks | **15+** | Automated lifecycle triggers — quality gates, session memory, cost tracking |
|
|
155
|
-
| Rules | **15** | Language-specific coding standards |
|
|
151
|
+
See `skills/agent-handoff/SKILL.md` for the full handoff protocol and standard chains.
|
|
156
152
|
|
|
157
153
|
---
|
|
158
154
|
|
|
159
|
-
##
|
|
155
|
+
## Kodelyth Memory — Local Self-Learning (new in v1.4.0)
|
|
160
156
|
|
|
161
|
-
|
|
162
|
-
|
|
163
|
-
### Kodelyth Exclusives
|
|
164
|
-
|
|
165
|
-
### `kodelyth-advisor` — The Master Guide
|
|
166
|
-
A decade-seasoned principal engineer who reads your emotional state, understands what you're struggling with, and gives you the exact tool or approach needed. No hand-holding, no noise — direct senior-grade direction.
|
|
157
|
+
The 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.
|
|
167
158
|
|
|
168
159
|
```
|
|
169
|
-
|
|
160
|
+
Session 1 (March):
|
|
161
|
+
You: "Stripe webhook signatures are failing in production"
|
|
162
|
+
AI: [helps you debug, you discover raw body parser is required]
|
|
163
|
+
You: "Perfect, that worked, thanks"
|
|
164
|
+
→ ECC's Stop hook queues the lesson for review.
|
|
165
|
+
→ You confirm with /memory review-pending → stored locally.
|
|
166
|
+
|
|
167
|
+
Session 2 (August, new project):
|
|
168
|
+
You: "I need to add Stripe webhooks to checkout"
|
|
169
|
+
AI: I checked your memory — you solved this exact problem in March.
|
|
170
|
+
Raw body parser before signature validation, test with stripe-cli
|
|
171
|
+
not curl. Want me to apply the same pattern?
|
|
172
|
+
You: "Yes, do it"
|
|
173
|
+
→ 30 minutes of past debugging saved in 3 seconds.
|
|
170
174
|
```
|
|
171
175
|
|
|
172
|
-
###
|
|
173
|
-
Never guesses. Traces every bug to its exact origin through a structured hypothesis protocol. Covers JS/TS, Python, Go, Rust, SQL — with advanced techniques like binary search debug, git bisect, and minimal reproducers.
|
|
176
|
+
### How it works
|
|
174
177
|
|
|
175
|
-
|
|
176
|
-
|
|
177
|
-
|
|
178
|
+
| Layer | Mechanism |
|
|
179
|
+
|---|---|
|
|
180
|
+
| **Capture** | A Stop hook scans your session JSONL, extracts (problem, approach, gotchas, tags). Queues for your review — never auto-stores. |
|
|
181
|
+
| **Storage** | `~/.kodelyth/memory/memories.jsonl` — append-only log on your disk only. Override with `KODELYTH_MEMORY_DIR`. |
|
|
182
|
+
| **Retrieval** | BM25 keyword + tag search. Pure JS, sub-millisecond, no embeddings, no network. |
|
|
183
|
+
| **Injection** | A SessionStart hook builds a cache-friendly context block: stable prefix (your patterns + recent project memories) → variable suffix (relevant to current task). |
|
|
184
|
+
| **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. |
|
|
178
185
|
|
|
179
|
-
###
|
|
180
|
-
Reviews frontend code for usability, interaction logic, and WCAG 2.1 AA accessibility. Never touches design aesthetic — only behavior. Catches what real users will feel before they feel it.
|
|
186
|
+
### Privacy
|
|
181
187
|
|
|
182
|
-
|
|
183
|
-
use ux-reviewer
|
|
184
|
-
```
|
|
188
|
+
Every 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).
|
|
185
189
|
|
|
186
|
-
###
|
|
187
|
-
Detects breaking API changes before they ship. Classifies every change as BREAKING / SAFE / WARN, generates a consumer impact report, and enforces versioning, deprecation headers, and idempotency.
|
|
190
|
+
### Slash command
|
|
188
191
|
|
|
189
192
|
```
|
|
190
|
-
|
|
193
|
+
/memory # Stats and recent memories
|
|
194
|
+
/memory recall <query> # BM25 search
|
|
195
|
+
/memory remember "<title>" # Capture (interactive — confirms before storing)
|
|
196
|
+
/memory review-pending # Review Stop-hook candidates
|
|
197
|
+
/memory forget <id> # Soft-delete a memory
|
|
198
|
+
/memory inject [--query <text>] # Print what your AI sees about you
|
|
191
199
|
```
|
|
192
200
|
|
|
193
|
-
###
|
|
194
|
-
The engineer who sits next to you before you start coding. Catches the wrong approach, over-engineered solutions, and missed edge cases — before a single line is written.
|
|
195
|
-
|
|
196
|
-
```
|
|
197
|
-
use pair-programmer
|
|
198
|
-
```
|
|
201
|
+
### Honest limits
|
|
199
202
|
|
|
200
|
-
|
|
201
|
-
|
|
203
|
+
- It is **not** model fine-tuning. The LLM never changes. We give it smarter context.
|
|
204
|
+
- Cache savings apply to Anthropic and OpenAI. Other models (Gemini, Llama, Mistral) get the recall quality without the discount.
|
|
205
|
+
- 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.
|
|
202
206
|
|
|
203
|
-
|
|
204
|
-
use migration-guide
|
|
205
|
-
```
|
|
207
|
+
See `skills/kodelyth-memory/SKILL.md` for the full design + CLI reference.
|
|
206
208
|
|
|
207
209
|
---
|
|
208
210
|
|
|
209
|
-
|
|
211
|
+
## What's Inside
|
|
210
212
|
|
|
211
|
-
|
|
|
212
|
-
|
|
213
|
-
|
|
|
214
|
-
|
|
|
215
|
-
|
|
|
216
|
-
|
|
|
217
|
-
|
|
|
218
|
-
|
|
|
219
|
-
| Quality | `refactor-cleaner`, `code-simplifier`, `performance-optimizer`, `type-design-analyzer` |
|
|
220
|
-
| Testing | `tdd-guide`, `e2e-runner`, `pr-test-analyzer` |
|
|
221
|
-
| Documentation | `doc-updater`, `docs-lookup`, `comment-analyzer` |
|
|
222
|
-
| Open Source | `opensource-forker`, `opensource-packager`, `opensource-sanitizer` |
|
|
213
|
+
| Component | Count | Description |
|
|
214
|
+
|---|---|---|
|
|
215
|
+
| Agents | **59** | Specialist subagents — reviewers, planners, debuggers, architects, doctors, memory curator |
|
|
216
|
+
| Skills | **188** | Domain knowledge — patterns, testing, security, DevOps, intent routing, memory |
|
|
217
|
+
| Commands | **80** | Slash command workflows (`/tdd`, `/plan`, `/memory`, etc.) |
|
|
218
|
+
| Hooks | **18+** | Quality gates, secret scanning, branch checks, memory inject + capture |
|
|
219
|
+
| Rules | **16** | Always-on coding standards + intent routing + memory protocol |
|
|
220
|
+
| Memory | **local** | BM25-indexed personal memory at `~/.kodelyth/memory/` (zero deps) |
|
|
223
221
|
|
|
224
222
|
---
|
|
225
223
|
|
|
226
|
-
##
|
|
224
|
+
## Agent Arsenal
|
|
227
225
|
|
|
228
|
-
|
|
226
|
+
### Kodelyth Exclusives — The 13 Agents That Make ECC
|
|
229
227
|
|
|
230
|
-
|
|
231
|
-
|
|
232
|
-
|
|
233
|
-
|
|
234
|
-
|
|
235
|
-
|
|
228
|
+
| Agent | One-line job |
|
|
229
|
+
|---|---|
|
|
230
|
+
| `kodelyth-advisor` | Master guide — picks the right tool when you don't know where to start |
|
|
231
|
+
| `kodelyth-memory` | Curates your local memory — recalls past solutions, captures new ones |
|
|
232
|
+
| `pair-programmer` | The engineer who sits next to you **before** you write the code |
|
|
233
|
+
| `debug-detective` | Never guesses — traces every bug to root cause |
|
|
234
|
+
| `silent-failure-hunter` | Finds bugs that don't throw errors |
|
|
235
|
+
| `ux-reviewer` | Reviews UX behavior + WCAG 2.1 AA accessibility (never touches design) |
|
|
236
|
+
| `api-guardian` | Detects breaking API changes before they ship |
|
|
237
|
+
| `migration-guide` | Framework / language version upgrades, phase by phase |
|
|
238
|
+
| `dependency-doctor` | npm/pip/cargo/maven hell — CVE triage, lockfile diagnosis |
|
|
239
|
+
| `git-rescue` | Recovers from broken git states without destroying history |
|
|
240
|
+
| `release-captain` | Owns the release ritual — semver, tagging, publishing, rollback |
|
|
241
|
+
| `env-debugger` | "Works on my machine" hunter — env, config, secrets, layers |
|
|
242
|
+
| `flake-hunter` | Stabilizes flaky tests — never adds blind retries |
|
|
236
243
|
|
|
237
|
-
|
|
244
|
+
### Agent Categories
|
|
238
245
|
|
|
239
|
-
|
|
|
246
|
+
| Category | Agents |
|
|
240
247
|
|---|---|
|
|
241
|
-
|
|
|
242
|
-
|
|
|
243
|
-
|
|
|
244
|
-
|
|
|
245
|
-
|
|
|
246
|
-
|
|
|
247
|
-
|
|
|
248
|
-
|
|
|
249
|
-
|
|
|
250
|
-
|
|
|
251
|
-
|
|
|
252
|
-
|
|
|
253
|
-
|
|
254
|
-
Intent detection runs across all platforms — Claude Code, Windsurf, Cursor, Codex — anywhere Kodelyth Lens can read conversation logs. The dashboard shows both explicit and auto-detected agent activations in the leaderboard.
|
|
248
|
+
| **Guidance** | `kodelyth-advisor`, `pair-programmer`, `planner`, `architect`, `code-architect`, `chief-of-staff`, `migration-guide` |
|
|
249
|
+
| **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` |
|
|
250
|
+
| **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` |
|
|
251
|
+
| **Debugging** | `debug-detective`, `silent-failure-hunter`, `flake-hunter` |
|
|
252
|
+
| **Security & API** | `security-reviewer`, `api-guardian` |
|
|
253
|
+
| **Performance** | `performance-optimizer` |
|
|
254
|
+
| **Quality** | `refactor-cleaner`, `code-simplifier`, `type-design-analyzer` |
|
|
255
|
+
| **Testing** | `tdd-guide`, `e2e-runner`, `pr-test-analyzer`, `flake-hunter` |
|
|
256
|
+
| **Documentation** | `doc-updater`, `docs-lookup`, `comment-analyzer` |
|
|
257
|
+
| **Release & Ops** | `release-captain`, `git-rescue` |
|
|
258
|
+
| **Open Source** | `opensource-forker`, `opensource-packager`, `opensource-sanitizer` |
|
|
259
|
+
| **Specialized** | `seo-specialist`, `ux-reviewer`, `code-explorer` |
|
|
255
260
|
|
|
256
261
|
---
|
|
257
262
|
|
|
258
263
|
## Hooks — Always Running in Background
|
|
259
264
|
|
|
260
|
-
Once installed, these hooks run automatically with zero configuration:
|
|
265
|
+
Once installed (Claude Code target), these hooks run automatically with zero configuration:
|
|
261
266
|
|
|
262
267
|
| Hook | What it does |
|
|
263
268
|
|---|---|
|
|
@@ -266,8 +271,7 @@ Once installed, these hooks run automatically with zero configuration:
|
|
|
266
271
|
| Quality gate | Runs type checks and formatting after edits |
|
|
267
272
|
| Git push reminder | Prompts review before pushing |
|
|
268
273
|
| Config protection | Blocks weakening of linter/formatter configs |
|
|
269
|
-
|
|
|
270
|
-
| Desktop notify | macOS notification when Claude finishes a long task |
|
|
274
|
+
| Desktop notify | macOS notification when a long task finishes |
|
|
271
275
|
| MCP health check | Validates MCP servers before calling them |
|
|
272
276
|
| Test reminder | Prompts to write tests when code is edited without tests |
|
|
273
277
|
| Smart suggest | After each response, suggests the next logical agent |
|
|
@@ -278,11 +282,13 @@ Once installed, these hooks run automatically with zero configuration:
|
|
|
278
282
|
## Usage After Install
|
|
279
283
|
|
|
280
284
|
### Start here (new users)
|
|
285
|
+
|
|
281
286
|
```
|
|
282
287
|
/kodelyth-quickstart
|
|
283
288
|
```
|
|
284
289
|
|
|
285
|
-
### Core workflows
|
|
290
|
+
### Core workflows — explicit invocation
|
|
291
|
+
|
|
286
292
|
```bash
|
|
287
293
|
use kodelyth-advisor # Not sure what to do? Start here
|
|
288
294
|
use pair-programmer # Think through approach before writing code
|
|
@@ -295,9 +301,19 @@ use ux-reviewer # Review frontend UX and accessibility
|
|
|
295
301
|
use migration-guide # Upgrading a framework or language version
|
|
296
302
|
use tdd-guide # Write tests the right way
|
|
297
303
|
use performance-optimizer # Diagnose and fix slowness
|
|
304
|
+
use dependency-doctor # npm/pip/cargo dep hell
|
|
305
|
+
use git-rescue # Broken git state, lost commits, bad rebase
|
|
306
|
+
use release-captain # Cut a clean release with rollback plan
|
|
307
|
+
use env-debugger # "Works on my machine" — env/config/secrets
|
|
308
|
+
use flake-hunter # Stabilize flaky tests
|
|
298
309
|
```
|
|
299
310
|
|
|
300
|
-
###
|
|
311
|
+
### Or just describe your problem
|
|
312
|
+
|
|
313
|
+
The 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.
|
|
314
|
+
|
|
315
|
+
### Load language patterns (skills)
|
|
316
|
+
|
|
301
317
|
```bash
|
|
302
318
|
/typescript-patterns # TypeScript + React + Next.js
|
|
303
319
|
/python-patterns # Python best practices
|
|
@@ -308,54 +324,55 @@ use performance-optimizer # Diagnose and fix slowness
|
|
|
308
324
|
```
|
|
309
325
|
|
|
310
326
|
### Advanced skills
|
|
327
|
+
|
|
311
328
|
```bash
|
|
312
329
|
/git-mastery # Trunk-based dev, rebase, bisect, monorepos
|
|
313
330
|
/observability # Structured logging, metrics, OpenTelemetry, SLOs
|
|
314
331
|
/smart-debug # Systematic debugging framework
|
|
332
|
+
/intent-routing # How the auto-routing system works
|
|
333
|
+
/agent-handoff # Standard multi-agent handoff chains
|
|
315
334
|
```
|
|
316
335
|
|
|
317
336
|
---
|
|
318
337
|
|
|
319
338
|
## What Gets Installed
|
|
320
339
|
|
|
321
|
-
### Claude Code
|
|
340
|
+
### Claude Code — `~/.claude/`
|
|
322
341
|
|
|
323
342
|
| Source | Destination | What it does |
|
|
324
343
|
|---|---|---|
|
|
325
|
-
| `agents/` | `~/.claude/agents/` | All
|
|
326
|
-
| `skills/` | `~/.claude/skills/` | All
|
|
327
|
-
| `hooks/hooks.json` | `~/.claude/hooks/` | Automated quality gates
|
|
328
|
-
| `rules/` | `~/.claude/rules/` | Always-on
|
|
344
|
+
| `agents/` | `~/.claude/agents/` | All 59 subagents available globally |
|
|
345
|
+
| `skills/` | `~/.claude/skills/` | All 188 skills loadable via commands |
|
|
346
|
+
| `hooks/hooks.json` | `~/.claude/hooks/` | Automated quality gates |
|
|
347
|
+
| `rules/` | `~/.claude/rules/` | Always-on standards + intent routing |
|
|
329
348
|
| `commands/` | `~/.claude/commands/` | Slash commands (`/tdd`, `/plan`, etc.) |
|
|
330
|
-
| `dashboard/` | `~/.kodelyth/lens/` | Kodelyth Lens usage dashboard |
|
|
331
349
|
|
|
332
|
-
### Windsurf
|
|
350
|
+
### Windsurf — `.windsurf/` (project) or `~/.codeium/windsurf/` (global)
|
|
333
351
|
|
|
334
352
|
| Source | Destination | Notes |
|
|
335
353
|
|---|---|---|
|
|
336
354
|
| `agents/` | `.windsurf/agents/` | Call with `use <agent-name>` in Cascade |
|
|
337
355
|
| `skills/` | `.windsurf/skills/` | Domain knowledge, loadable via chat |
|
|
338
356
|
| `rules/` | `.windsurf/rules/` | Flattened rule files |
|
|
339
|
-
| `rules/common/` | `.windsurfrules` | Concatenated — Windsurf reads this automatically |
|
|
340
|
-
|
|
341
|
-
The `.windsurfrules` file is auto-generated from all common rules and loaded by Windsurf Cascade on every session. No manual setup needed.
|
|
357
|
+
| `rules/common/` | `.windsurfrules` | Concatenated — Windsurf reads this automatically every session |
|
|
342
358
|
|
|
343
|
-
###
|
|
359
|
+
### Cursor — `.cursor/` in project root
|
|
344
360
|
|
|
345
|
-
| Source | Destination |
|
|
346
|
-
|
|
347
|
-
| `
|
|
348
|
-
| `skills/` |
|
|
349
|
-
| `commands/` | `~/.codex/commands/` | Slash commands |
|
|
350
|
-
| `rules/` | `~/.codex/rules/` | Coding standards (flattened) |
|
|
361
|
+
| Source | Destination |
|
|
362
|
+
|---|---|
|
|
363
|
+
| `rules/` | `.cursor/rules/` |
|
|
364
|
+
| `skills/` | `.cursor/skills/` |
|
|
351
365
|
|
|
352
|
-
###
|
|
366
|
+
### Codex CLI — `~/.codex/`
|
|
353
367
|
|
|
354
|
-
| Source | Destination |
|
|
355
|
-
|
|
356
|
-
| `
|
|
368
|
+
| Source | Destination |
|
|
369
|
+
|---|---|
|
|
370
|
+
| `agents/` | `~/.codex/agents/` |
|
|
371
|
+
| `skills/` | `~/.codex/skills/` |
|
|
372
|
+
| `commands/` | `~/.codex/commands/` |
|
|
373
|
+
| `rules/` | `~/.codex/rules/` |
|
|
357
374
|
|
|
358
|
-
### Antigravity
|
|
375
|
+
### Antigravity — `.agent/` in project root
|
|
359
376
|
|
|
360
377
|
| Source | Destination |
|
|
361
378
|
|---|---|
|
|
@@ -363,6 +380,12 @@ The `.windsurfrules` file is auto-generated from all common rules and loaded by
|
|
|
363
380
|
| `commands/` | `.agent/workflows/` |
|
|
364
381
|
| `rules/` | `.agent/rules/` |
|
|
365
382
|
|
|
383
|
+
### OpenCode — `.opencode/` in project root
|
|
384
|
+
|
|
385
|
+
| Source | Destination |
|
|
386
|
+
|---|---|
|
|
387
|
+
| `rules/` | `.opencode/rules/` |
|
|
388
|
+
|
|
366
389
|
---
|
|
367
390
|
|
|
368
391
|
## Install Profiles — Language Bundles
|
|
@@ -370,15 +393,9 @@ The `.windsurfrules` file is auto-generated from all common rules and loaded by
|
|
|
370
393
|
Works with any target:
|
|
371
394
|
|
|
372
395
|
```bash
|
|
373
|
-
# Claude Code
|
|
374
|
-
|
|
375
|
-
npx
|
|
376
|
-
|
|
377
|
-
# Windsurf
|
|
378
|
-
npx github:sifxprime/kodelyth-ecc --target windsurf-project --profile nextjs
|
|
379
|
-
|
|
380
|
-
# Antigravity
|
|
381
|
-
npx github:sifxprime/kodelyth-ecc --target antigravity --profile python-api
|
|
396
|
+
npx kodelyth-ecc --profile nextjs # Claude Code
|
|
397
|
+
npx kodelyth-ecc --target windsurf-project --profile nextjs # Windsurf
|
|
398
|
+
npx kodelyth-ecc --target antigravity --profile python-api # Antigravity
|
|
382
399
|
```
|
|
383
400
|
|
|
384
401
|
Available profiles:
|
|
@@ -391,9 +408,10 @@ Available profiles:
|
|
|
391
408
|
| `mobile` | Kotlin + Swift rules |
|
|
392
409
|
| `backend` | Go + Python + Java rules |
|
|
393
410
|
|
|
394
|
-
|
|
411
|
+
Or specify languages directly:
|
|
412
|
+
|
|
395
413
|
```bash
|
|
396
|
-
npx
|
|
414
|
+
npx kodelyth-ecc typescript python golang rust java kotlin php swift cpp dart
|
|
397
415
|
```
|
|
398
416
|
|
|
399
417
|
---
|
|
@@ -404,21 +422,55 @@ npx github:sifxprime/kodelyth-ecc typescript python golang rust java kotlin php
|
|
|
404
422
|
|---|---|---|---|
|
|
405
423
|
| Claude Code | Full | `claude-home` (default) | Agents, skills, commands, hooks, rules |
|
|
406
424
|
| Windsurf | Full | `windsurf-project` / `windsurf-home` | Agents, skills, rules, `.windsurfrules` |
|
|
407
|
-
| Google Antigravity | Full | `antigravity` | Agents → skills, commands → workflows, rules |
|
|
408
425
|
| Cursor | Full | `cursor-project` | Rules, skills |
|
|
409
426
|
| Codex CLI | Full | `codex-home` | Agents, skills, commands, rules |
|
|
427
|
+
| Google Antigravity | Full | `antigravity` | Agents → skills, commands → workflows, rules |
|
|
410
428
|
| OpenCode | Full | `opencode` | Rules |
|
|
411
429
|
|
|
430
|
+
**OS support:** macOS, Linux (`install.sh`), Windows (`install.ps1`), or any OS with Node.js 18+ (`npx`).
|
|
431
|
+
|
|
432
|
+
---
|
|
433
|
+
|
|
434
|
+
## Privacy & Philosophy
|
|
435
|
+
|
|
436
|
+
ECC is **100% local files**. No telemetry, no cloud, no account, no tracking. Everything is markdown your AI reads on every session.
|
|
437
|
+
|
|
438
|
+
- **No dashboard** — we removed it in v1.3.0 because it wasn't giving accurate cross-platform data and added confusion. The toolkit is now leaner.
|
|
439
|
+
- **No phoning home** — install scripts copy files and exit. That's it.
|
|
440
|
+
- **You own everything** — fork the repo, edit any agent, write your own.
|
|
441
|
+
|
|
412
442
|
---
|
|
413
443
|
|
|
414
444
|
## Contributing
|
|
415
445
|
|
|
416
|
-
We welcome new agents, skills, and
|
|
446
|
+
We welcome new agents, skills, hooks, and intent routing patterns that meet the **Kodelyth Standard** — specialist personas, model-agnostic, production-grade examples.
|
|
447
|
+
|
|
448
|
+
When 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.
|
|
417
449
|
|
|
418
450
|
See [CONTRIBUTING.md](CONTRIBUTING.md) for templates, checklists, and the full Kodelyth Standard.
|
|
419
451
|
|
|
420
452
|
---
|
|
421
453
|
|
|
454
|
+
## Changelog
|
|
455
|
+
|
|
456
|
+
See [CHANGELOG.md](CHANGELOG.md). v1.4.0 highlights:
|
|
457
|
+
|
|
458
|
+
- **Added:** `kodelyth-memory` — local self-learning memory (BM25, zero deps, cache-friendly)
|
|
459
|
+
- **Added:** SessionStart hook injects relevant past memories; Stop hook queues candidates for review
|
|
460
|
+
- **Added:** `/memory` slash command + `kodelyth-memory` agent + `memory-protocol` rule
|
|
461
|
+
- **Added:** 12 new tests (38 total passing)
|
|
462
|
+
- **Honest disclosure:** "Self-learning" means smarter context retrieval, not model fine-tuning
|
|
463
|
+
|
|
464
|
+
v1.3.0 highlights:
|
|
465
|
+
|
|
466
|
+
- **Removed:** Lens dashboard (was inaccurate across cloud-AI platforms)
|
|
467
|
+
- **Added:** God-tier `agent-intent-routing` rule — 10 priority tiers, 50+ patterns
|
|
468
|
+
- **Added:** 5 new agents — `dependency-doctor`, `git-rescue`, `release-captain`, `env-debugger`, `flake-hunter`
|
|
469
|
+
- **Added:** 2 new skills — `intent-routing` (routing system docs), `agent-handoff` (chain protocol)
|
|
470
|
+
- **Improved:** Cross-platform install parity — full Windows support for all targets
|
|
471
|
+
|
|
472
|
+
---
|
|
473
|
+
|
|
422
474
|
<div align="center">
|
|
423
475
|
|
|
424
476
|
Built and maintained by **Kodelyth** — [@sifxprime](https://github.com/sifxprime)
|
package/VERSION
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
1.
|
|
1
|
+
1.4.0
|