@xelth/eck-snapshot 5.4.1 → 5.4.3

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.
Files changed (2) hide show
  1. package/README.md +92 -63
  2. package/package.json +1 -1
package/README.md CHANGED
@@ -1,106 +1,135 @@
1
+ # eck-snapshot
1
2
 
2
- # eckSnapshot (v5.4.0)
3
+ A CLI tool that packs your entire Git repository into a single text file optimized for LLMs. Give any AI full project context in one copy-paste.
3
4
 
4
- A specialized CLI tool designed to create and restore single-file text snapshots of Git repositories. It is specifically optimized for providing full project context to Large Language Models (LLMs) like Claude, Gemini, and OpenCode.
5
+ ```bash
6
+ npm install -g @xelth/eck-snapshot
7
+ ```
5
8
 
6
- 🎉 **WE ARE BACK ON NPM!** Version 5.4.0 and onwards are officially available via the npm registry.
9
+ ## Recommended AI Setup
7
10
 
8
- ## 🚀 Quick Start
11
+ For best results, we recommend splitting roles between models:
9
12
 
10
- ```bash
11
- # Install globally via npm
12
- npm install -g @xelth/eck-snapshot
13
+ - **Architect** (large context window): Gemini, Grok Fast, ChatGPT — upload the full snapshot, design the architecture, plan tasks
14
+ - **Coder** (execution): Claude (via Claude Code), GLM (via OpenCode) — receive tasks from the architect, write and fix code
13
15
 
14
- # Create a snapshot of your current project
16
+ eck-snapshot generates tailored instructions (`CLAUDE.md`, `AGENTS.md`) for each role automatically.
17
+
18
+ ## Core Workflow
19
+
20
+ ### 1. Full Snapshot
21
+
22
+ Run `eck-snapshot` in your project root. It scans every tracked file, filters out noise (lock files, build artifacts, secrets), and produces a single `.md` file ready for an AI chat.
23
+
24
+ ```bash
15
25
  eck-snapshot
26
+ # -> .eck/snapshots/eckMyProject_26-02-15_12-00_abc1234.md
16
27
  ```
17
28
 
18
- ## Core Features
29
+ Upload the file to your architect AI and start working.
19
30
 
20
- - **Skeleton Mode:** Strips function bodies using Tree-sitter and Babel to save massive amounts of context tokens.
21
- - **Delta Updates:** Tracks changes via Git anchors and generates incremental snapshots (`eck-snapshot update`).
22
- - **Royal Court Architecture:** Multi-agent protocol with dedicated modes for Claude Sonnet (JAS), Claude Opus (JAO), and Gemini (JAG).
23
- - **GLM Z.AI Worker Fleet:** Built-in MCP server integration for delegating heavy coding tasks to specialized AI workers.
24
- - **Security:** Built-in `SecretScanner` automatically redacts API keys and sensitive credentials before they hit the LLM context.
25
- - **Context Profiles:** Smart filtering using auto-detected or manual profiles (e.g., `--profile backend`).
31
+ ### 2. Incremental Update
26
32
 
27
- ---
33
+ After you make changes, don't re-send the entire project. Send only what changed since the last full snapshot:
28
34
 
29
- ## 🛠️ The Core Workflow
30
-
31
- ### 1. Initial Context (Maximum Compression)
32
- Create a lightweight map of your entire project. Bodies of functions are hidden, allowing huge monoliths to fit into the AI's context window.
33
35
  ```bash
34
- eck-snapshot --skeleton
35
- # -> Generates: .eck/snapshots/eck[Name]_[Hash]_sk.md
36
+ eck-snapshot update
37
+ # -> .eck/snapshots/eckMyProject_26-02-15_14-30_abc1234_up1.md
36
38
  ```
37
39
 
38
- ### 2. Lazy Loading (On-Demand Details)
39
- If the AI needs to see the exact implementation of specific files, it can request them on demand.
40
+ This uses a Git anchor (saved automatically during full snapshot) to detect all modified files and includes their full content. No redundant diffs, no wasted tokens.
41
+
42
+ ## Context Profiles
43
+
44
+ Large repositories waste tokens on irrelevant code. Profiles let you partition the codebase so the AI only sees what matters.
45
+
46
+ ### Auto-Detection
47
+
48
+ Let AI scan your directory tree and generate profiles automatically:
49
+
40
50
  ```bash
41
- eck-snapshot show src/auth.js src/utils/hash.js
51
+ eck-snapshot profile-detect
52
+ # -> Saves profiles to .eck/profiles.json
42
53
  ```
43
54
 
44
- ### 3. Incremental Updates (Delta)
45
- As you apply changes, the AI loses context. Instead of re-sending the full repository, send only what changed since the last snapshot!
55
+ ### Manual Guide
56
+
57
+ For very large repos where auto-detection is too slow, generate a prompt guide, paste it into a powerful Web LLM (Gemini, ChatGPT), and save the resulting JSON:
58
+
46
59
  ```bash
47
- eck-snapshot update
48
- # -> Generates an update snapshot with git diffs and modified files
60
+ eck-snapshot generate-profile-guide
61
+ # -> .eck/profile_generation_guide.md (paste into AI, get profiles back)
49
62
  ```
50
63
 
51
- ---
64
+ ### Using Profiles
52
65
 
53
- ## 👑 Royal Court Architecture & GLM Z.AI
66
+ ```bash
67
+ eck-snapshot --profile # List all available profiles
68
+ eck-snapshot --profile backend # Use a named profile
69
+ eck-snapshot --profile backend --skeleton # Profile + skeleton mode
70
+ eck-snapshot --profile "src/**/*.rs,-**/test_*" # Ad-hoc glob filtering
71
+ ```
54
72
 
55
- `eck-snapshot` is designed to orchestrate a hierarchy of AI agents:
73
+ Profiles work with both full snapshots and incremental updates.
56
74
 
57
- - **Senior Architect:** (You / Gemini / ChatGPT) - Directs the high-level strategy.
58
- - **Junior Architects:**
59
- - `JAS` (Sonnet 4.5): Fast manager for standard features. Run `eck-snapshot --jas`.
60
- - `JAO` (Opus 4.5): Deep thinker for critical architecture. Run `eck-snapshot --jao`.
61
- - `JAG` (Gemini 3 Pro): Massive context handler. Run `eck-snapshot --jag`.
75
+ ## Smart Filtering
62
76
 
63
- ### MCP Server Integration
64
- Delegate heavy coding tasks (>100 lines) to the **GLM Z.AI Worker Fleet** to save expensive context window tokens.
77
+ eck-snapshot automatically detects your project type (Rust, Node.js, Android, Python, etc.) and excludes language-specific noise:
65
78
 
66
- 1. Get your API key from [Z.AI](https://z.ai) and export it: `export ZAI_API_KEY="your-key-here"`
67
- 2. Setup the MCP servers for Claude Code or OpenCode:
68
- ```bash
69
- eck-snapshot setup-mcp --both
70
- ```
71
- 3. Your AI will now have access to specialized tools: `glm_zai_frontend`, `glm_zai_backend`, `glm_zai_qa`, `glm_zai_refactor`, and the `eck_finish_task` commit tool.
79
+ - **Rust**: `Cargo.lock`, `target/`
80
+ - **Node.js**: `package-lock.json`, `node_modules/`
81
+ - **Android**: build artifacts, generated code
82
+ - **All projects**: `.git/`, IDE configs, binary files
72
83
 
73
- ---
84
+ The built-in `SecretScanner` also redacts API keys, tokens, and credentials before they reach the AI.
74
85
 
75
- ## 🧩 Context Profiles
86
+ ## Multi-Agent Architecture
76
87
 
77
- If your repository is huge, you can partition it using Context Profiles:
88
+ eck-snapshot generates tailored `CLAUDE.md` instructions for different AI agent roles:
78
89
 
79
90
  ```bash
80
- # Auto-detect profiles using AI
81
- eck-snapshot profile-detect
91
+ eck-snapshot --jas # Junior Architect Sonnet - fast, standard features
92
+ eck-snapshot --jao # Junior Architect Opus - deep, critical architecture
93
+ eck-snapshot --jag # Junior Architect Gemini - massive context tasks
94
+ ```
82
95
 
83
- # List available profiles
84
- eck-snapshot --profile
96
+ ### MCP Server Integration
85
97
 
86
- # Use a specific profile
87
- eck-snapshot --profile backend
98
+ Delegate coding tasks to the GLM Z.AI Worker Fleet via MCP:
88
99
 
89
- # Ad-hoc inclusion/exclusion
90
- eck-snapshot --profile "src/**/*.js,-**/*.test.js"
100
+ ```bash
101
+ export ZAI_API_KEY="your-key"
102
+ eck-snapshot setup-mcp --both # Setup for Claude Code + OpenCode
91
103
  ```
92
104
 
93
- ## 🔐 Environment Syncing
105
+ This gives your AI access to specialized workers: `glm_zai_frontend`, `glm_zai_backend`, `glm_zai_qa`, `glm_zai_refactor`, and the `eck_finish_task` commit tool.
106
+
107
+ ## Skeleton Mode & Lazy Loading
94
108
 
95
- Securely share your `.eck/` configuration (profiles, roadmap, AI instructions) between machines without committing them to the public git history:
109
+ For extremely large projects, skeleton mode strips function bodies and keeps only signatures, types, and structure:
96
110
 
97
111
  ```bash
98
- # Encrypt and pack .eck/ config files
99
- eck-snapshot env push
112
+ eck-snapshot --skeleton
113
+ ```
114
+
115
+ When using skeleton mode, the AI can request full content of specific files on demand:
116
+
117
+ ```bash
118
+ eck-snapshot show src/auth.rs src/handlers/sync.rs
119
+ ```
120
+
121
+ Useful for initial orientation in massive codebases, but full snapshots with profiles are usually more practical.
122
+
123
+ ## Other Commands
100
124
 
101
- # Decrypt and restore on another machine
102
- eck-snapshot env pull
125
+ ```bash
126
+ eck-snapshot restore <snapshot> # Restore files from a snapshot to disk
127
+ eck-snapshot prune <snapshot> # AI-powered snapshot size reduction
128
+ eck-snapshot doctor # Check project health
129
+ eck-snapshot env push # Encrypt and sync .eck/ config between machines
130
+ eck-snapshot env pull # Restore .eck/ config on another machine
103
131
  ```
104
132
 
105
133
  ## License
106
- MIT License
134
+
135
+ MIT
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@xelth/eck-snapshot",
3
- "version": "5.4.1",
3
+ "version": "5.4.3",
4
4
  "description": "A powerful CLI tool to create and restore single-file text snapshots of Git repositories and directories. Optimized for AI context and LLM workflows.",
5
5
  "main": "index.js",
6
6
  "type": "module",