@rely-ai/caliber 1.24.1-dev.1773872765 → 1.24.1
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/README.md +64 -167
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -6,93 +6,56 @@
|
|
|
6
6
|
<a href="https://www.npmjs.com/package/@rely-ai/caliber"><img src="https://img.shields.io/npm/v/@rely-ai/caliber" alt="npm version"></a>
|
|
7
7
|
<a href="./LICENSE"><img src="https://img.shields.io/npm/l/@rely-ai/caliber" alt="license"></a>
|
|
8
8
|
<a href="https://nodejs.org"><img src="https://img.shields.io/node/v/@rely-ai/caliber" alt="node"></a>
|
|
9
|
-
<img src="https://img.shields.io/badge/caliber-89%2F100-green" alt="Caliber Score">
|
|
10
|
-
<img src="https://img.shields.io/badge/Claude_Code-supported-blue" alt="Claude Code">
|
|
11
|
-
<img src="https://img.shields.io/badge/Cursor-supported-blue" alt="Cursor">
|
|
12
|
-
<img src="https://img.shields.io/badge/Codex-supported-blue" alt="Codex">
|
|
13
9
|
</p>
|
|
14
10
|
|
|
15
|
-
---
|
|
16
|
-
|
|
17
|
-
### Try it — zero install, zero commitment
|
|
18
|
-
|
|
19
|
-
```bash
|
|
20
|
-
npx @rely-ai/caliber score
|
|
21
|
-
```
|
|
22
|
-
|
|
23
|
-
Score your AI agent config in 3 seconds. No API key. No changes to your code. Just a score.
|
|
24
|
-
|
|
25
|
-
> **Your code stays on your machine.** Scoring is 100% local — no LLM calls, no code sent anywhere. Generation uses your own AI subscription (Claude Code, Cursor) or your own API key (Anthropic, OpenAI, Vertex AI). Caliber never sees your code.
|
|
26
|
-
|
|
27
|
-
---
|
|
28
|
-
|
|
29
|
-
Caliber scores, generates, and keeps your AI agent configs in sync with your codebase. It fingerprints your project — languages, frameworks, dependencies, architecture — and produces tailored configs for **Claude Code**, **Cursor**, and **OpenAI Codex**. When your code evolves, Caliber detects the drift and updates your configs to match.
|
|
30
|
-
|
|
31
11
|
<p align="center">
|
|
32
12
|
<img src="assets/demo.gif" alt="Caliber demo" width="640">
|
|
33
13
|
</p>
|
|
34
14
|
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
Most repos start with a hand-written `CLAUDE.md` and nothing else. Here's what Caliber finds — and fixes:
|
|
38
|
-
|
|
39
|
-
```
|
|
40
|
-
Before After caliber init
|
|
41
|
-
────────────────────────────── ──────────────────────────────
|
|
42
|
-
|
|
43
|
-
Agent Config Score 35 / 100 Agent Config Score 94 / 100
|
|
44
|
-
Grade D Grade A
|
|
45
|
-
|
|
46
|
-
FILES & SETUP 6 / 25 FILES & SETUP 24 / 25
|
|
47
|
-
QUALITY 12 / 25 QUALITY 22 / 25
|
|
48
|
-
GROUNDING 7 / 20 GROUNDING 19 / 20
|
|
49
|
-
ACCURACY 5 / 15 ACCURACY 13 / 15
|
|
50
|
-
FRESHNESS 5 / 10 FRESHNESS 10 / 10
|
|
51
|
-
BONUS 0 / 5 BONUS 5 / 5
|
|
52
|
-
```
|
|
15
|
+
---
|
|
53
16
|
|
|
54
|
-
|
|
17
|
+
Caliber analyzes your codebase — languages, frameworks, dependencies, architecture — and generates tailored, high-quality configs for **Claude Code**, **Cursor**, and **OpenAI Codex**. If configs already exist, it scores them, fixes what's stale, and keeps everything in sync as your code evolves.
|
|
55
18
|
|
|
56
19
|
```bash
|
|
57
|
-
|
|
20
|
+
npm install -g @rely-ai/caliber
|
|
21
|
+
caliber init
|
|
58
22
|
```
|
|
59
23
|
|
|
60
|
-
|
|
61
|
-
|
|
62
|
-
Caliber never overwrites your existing configs without asking. The workflow mirrors code review:
|
|
24
|
+
No API key required — works with your existing **Claude Code** or **Cursor** subscription. Or bring your own key (Anthropic, OpenAI, Vertex AI, any OpenAI-compatible endpoint).
|
|
63
25
|
|
|
64
|
-
|
|
65
|
-
2. **Propose** — generate or improve configs, shown as a diff
|
|
66
|
-
3. **Review** — accept, refine via chat, or decline each change
|
|
67
|
-
4. **Backup** — originals saved to `.caliber/backups/` before every write
|
|
68
|
-
5. **Undo** — `caliber undo` restores everything to its previous state
|
|
26
|
+
## 💡 What Caliber Does
|
|
69
27
|
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
|
|
28
|
+
| Without Caliber | With Caliber |
|
|
29
|
+
|---|---|
|
|
30
|
+
| Hand-write CLAUDE.md, Cursor rules, AGENTS.md separately | One command generates all three, tuned to your actual codebase |
|
|
31
|
+
| Configs reference files that no longer exist | Deterministic scoring catches stale references and drift |
|
|
32
|
+
| New team members start with no AI context | `caliber init` gives any contributor a complete setup in seconds |
|
|
33
|
+
| Configs diverge across AI tools | Cross-platform parity — Claude, Cursor, and Codex stay consistent |
|
|
34
|
+
| No idea if your config is actually helping | Score your setup (A–F grade) and see exactly what to improve |
|
|
35
|
+
| AI keeps making the same mistakes | Session learning captures patterns and corrections automatically |
|
|
73
36
|
|
|
74
|
-
|
|
37
|
+
## ⚙️ How It Works
|
|
75
38
|
|
|
76
39
|
```
|
|
77
|
-
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
|
|
82
|
-
|
|
83
|
-
|
|
84
|
-
|
|
85
|
-
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
|
|
90
|
-
|
|
40
|
+
caliber init
|
|
41
|
+
│
|
|
42
|
+
├─ 1. 🔌 Setup Choose your LLM provider — Claude Code seat, Cursor seat,
|
|
43
|
+
│ or an API key (Anthropic, OpenAI, Vertex AI)
|
|
44
|
+
│
|
|
45
|
+
├─ 2. 🛠️ Engine Fingerprint your project, generate configs in parallel,
|
|
46
|
+
│ search community skills, and auto-refine against
|
|
47
|
+
│ deterministic scoring checks (up to 2 iterations)
|
|
48
|
+
│
|
|
49
|
+
├─ 3. 👀 Review See a diff of every proposed change — accept, refine
|
|
50
|
+
│ via chat, or decline. All originals are backed up
|
|
51
|
+
│
|
|
52
|
+
└─ 4. ✅ Finalize Write files, install auto-refresh hooks, and set up
|
|
53
|
+
session learning for continuous improvement
|
|
91
54
|
```
|
|
92
55
|
|
|
93
|
-
|
|
56
|
+
Already have a setup? If your existing config scores **95+**, Caliber skips full regeneration and applies targeted fixes to the specific checks that are failing.
|
|
94
57
|
|
|
95
|
-
### What It Generates
|
|
58
|
+
### 📦 What It Generates
|
|
96
59
|
|
|
97
60
|
**Claude Code**
|
|
98
61
|
- `CLAUDE.md` — Project context, build/test commands, architecture, conventions
|
|
@@ -110,18 +73,14 @@ Auto-refresh hooks run this loop automatically — on every commit or at the end
|
|
|
110
73
|
- `AGENTS.md` — Project context for Codex
|
|
111
74
|
- `.agents/skills/*/SKILL.md` — Skills for Codex
|
|
112
75
|
|
|
113
|
-
|
|
76
|
+
If these files already exist, Caliber audits them and suggests improvements — keeping what works, fixing what's stale, adding what's missing.
|
|
114
77
|
|
|
115
|
-
|
|
116
|
-
<summary><strong>Any Codebase</strong></summary>
|
|
78
|
+
## ✨ Key Features
|
|
117
79
|
|
|
80
|
+
### 🌍 Any Codebase
|
|
118
81
|
TypeScript, Python, Go, Rust, Java, Ruby, Terraform, and more. Language and framework detection is fully LLM-driven — no hardcoded mappings. Caliber works on any project.
|
|
119
82
|
|
|
120
|
-
|
|
121
|
-
|
|
122
|
-
<details>
|
|
123
|
-
<summary><strong>Any AI Tool</strong></summary>
|
|
124
|
-
|
|
83
|
+
### 🔀 Any AI Tool
|
|
125
84
|
Target a single platform or all three at once:
|
|
126
85
|
```bash
|
|
127
86
|
caliber init --agent claude # Claude Code only
|
|
@@ -131,26 +90,28 @@ caliber init --agent all # All three
|
|
|
131
90
|
caliber init --agent claude,cursor # Comma-separated
|
|
132
91
|
```
|
|
133
92
|
|
|
134
|
-
|
|
135
|
-
|
|
136
|
-
<details>
|
|
137
|
-
<summary><strong>Chat-Based Refinement</strong></summary>
|
|
138
|
-
|
|
93
|
+
### 💬 Chat-Based Refinement
|
|
139
94
|
Not happy with the generated output? During review, refine via natural language — describe what you want changed and Caliber iterates until you're satisfied.
|
|
140
95
|
|
|
141
|
-
|
|
142
|
-
|
|
143
|
-
<details>
|
|
144
|
-
<summary><strong>MCP Server Discovery</strong></summary>
|
|
145
|
-
|
|
96
|
+
### 🔗 MCP Server Discovery
|
|
146
97
|
Caliber detects the tools your project uses (databases, APIs, services) and auto-configures matching MCP servers for Claude Code and Cursor.
|
|
147
98
|
|
|
148
|
-
|
|
99
|
+
### 📊 Deterministic Scoring
|
|
100
|
+
`caliber score` evaluates your config quality without any LLM calls — purely by cross-referencing config files against your actual project filesystem.
|
|
149
101
|
|
|
150
|
-
|
|
151
|
-
|
|
102
|
+
```
|
|
103
|
+
Agent Config Score 88 / 100 Grade A
|
|
104
|
+
|
|
105
|
+
FILES & SETUP 17 / 17
|
|
106
|
+
QUALITY 21 / 23
|
|
107
|
+
GROUNDING 20 / 20
|
|
108
|
+
ACCURACY 10 / 15
|
|
109
|
+
FRESHNESS & SAFETY 10 / 10
|
|
110
|
+
BONUS 5 / 5
|
|
111
|
+
```
|
|
152
112
|
|
|
153
|
-
|
|
113
|
+
<details>
|
|
114
|
+
<summary>Scoring breakdown</summary>
|
|
154
115
|
|
|
155
116
|
| Category | Points | What it checks |
|
|
156
117
|
|---|---|---|
|
|
@@ -165,9 +126,7 @@ Every failing check includes structured fix data — when `caliber init` runs, t
|
|
|
165
126
|
|
|
166
127
|
</details>
|
|
167
128
|
|
|
168
|
-
|
|
169
|
-
<summary><strong>Session Learning</strong></summary>
|
|
170
|
-
|
|
129
|
+
### 🧠 Session Learning
|
|
171
130
|
Caliber watches your AI coding sessions and learns from them. Hooks capture tool usage, failures, and your corrections — then an LLM distills operational patterns into `CALIBER_LEARNINGS.md`.
|
|
172
131
|
|
|
173
132
|
```bash
|
|
@@ -177,13 +136,9 @@ caliber learn finalize # Manually trigger analysis (auto-runs on session end
|
|
|
177
136
|
caliber learn remove # Remove hooks
|
|
178
137
|
```
|
|
179
138
|
|
|
180
|
-
Learned items are categorized by type — **[correction]**, **[gotcha]**, **[fix]**, **[pattern]**, **[env]**, **[convention]** — and automatically deduplicated.
|
|
181
|
-
|
|
182
|
-
</details>
|
|
183
|
-
|
|
184
|
-
<details>
|
|
185
|
-
<summary><strong>Auto-Refresh</strong></summary>
|
|
139
|
+
Learned items are categorized by type — **[correction]**, **[gotcha]**, **[fix]**, **[pattern]**, **[env]**, **[convention]** — and automatically deduplicated. ROI tracking shows how much time and tokens the learnings save across sessions.
|
|
186
140
|
|
|
141
|
+
### 🔄 Auto-Refresh
|
|
187
142
|
Keep configs in sync with your codebase automatically:
|
|
188
143
|
|
|
189
144
|
| Hook | Trigger | What it does |
|
|
@@ -195,29 +150,24 @@ Keep configs in sync with your codebase automatically:
|
|
|
195
150
|
```bash
|
|
196
151
|
caliber hooks --install # Enable refresh hooks
|
|
197
152
|
caliber hooks --remove # Disable refresh hooks
|
|
153
|
+
caliber learn install # Enable learning hooks
|
|
198
154
|
```
|
|
199
155
|
|
|
200
|
-
The `refresh` command analyzes your git diff (committed, staged, and unstaged changes) and updates config files to reflect what changed.
|
|
201
|
-
|
|
202
|
-
</details>
|
|
203
|
-
|
|
204
|
-
<details>
|
|
205
|
-
<summary><strong>Fully Reversible</strong></summary>
|
|
156
|
+
The `refresh` command analyzes your git diff (committed, staged, and unstaged changes) and updates config files to reflect what changed. Works across multiple repos when run from a parent directory.
|
|
206
157
|
|
|
158
|
+
### 🛡️ Fully Reversible
|
|
159
|
+
Every change Caliber makes can be undone:
|
|
207
160
|
- **Automatic backups** — originals saved to `.caliber/backups/` before every write
|
|
208
161
|
- **Score regression guard** — if a regeneration produces a lower score, changes are auto-reverted
|
|
209
162
|
- **Full undo** — `caliber undo` restores everything to its previous state
|
|
210
163
|
- **Dry run** — preview changes with `--dry-run` before applying
|
|
211
164
|
|
|
212
|
-
|
|
213
|
-
|
|
214
|
-
## Commands
|
|
165
|
+
## 📋 Commands
|
|
215
166
|
|
|
216
167
|
| Command | Description |
|
|
217
168
|
|---|---|
|
|
218
|
-
| `caliber score` | Score config quality (deterministic, no LLM) |
|
|
219
|
-
| `caliber score --compare <ref>` | Compare current score against a git ref |
|
|
220
169
|
| `caliber init` | Full setup wizard — analyze, generate, review, install hooks |
|
|
170
|
+
| `caliber score` | Score config quality (deterministic, no LLM) |
|
|
221
171
|
| `caliber regenerate` | Re-analyze and regenerate configs (aliases: `regen`, `re`) |
|
|
222
172
|
| `caliber refresh` | Update docs based on recent code changes |
|
|
223
173
|
| `caliber skills` | Discover and install community skills |
|
|
@@ -227,60 +177,7 @@ The `refresh` command analyzes your git diff (committed, staged, and unstaged ch
|
|
|
227
177
|
| `caliber status` | Show current setup status |
|
|
228
178
|
| `caliber undo` | Revert all changes made by Caliber |
|
|
229
179
|
|
|
230
|
-
##
|
|
231
|
-
|
|
232
|
-
<details>
|
|
233
|
-
<summary><strong>Does it overwrite my existing configs?</strong></summary>
|
|
234
|
-
|
|
235
|
-
No. Caliber shows you a diff of every proposed change. You accept, refine, or decline each one. Originals are backed up automatically.
|
|
236
|
-
|
|
237
|
-
</details>
|
|
238
|
-
|
|
239
|
-
<details>
|
|
240
|
-
<summary><strong>Does it need an API key?</strong></summary>
|
|
241
|
-
|
|
242
|
-
**Scoring:** No. `caliber score` runs 100% locally with no LLM.
|
|
243
|
-
|
|
244
|
-
**Generation:** Uses your existing Claude Code or Cursor subscription (no API key needed), or bring your own key for Anthropic, OpenAI, or Vertex AI.
|
|
245
|
-
|
|
246
|
-
</details>
|
|
247
|
-
|
|
248
|
-
<details>
|
|
249
|
-
<summary><strong>What if I don't like what it generates?</strong></summary>
|
|
250
|
-
|
|
251
|
-
Refine it via chat during review, or decline the changes entirely. If you already accepted, `caliber undo` restores everything. You can also preview with `--dry-run`.
|
|
252
|
-
|
|
253
|
-
</details>
|
|
254
|
-
|
|
255
|
-
<details>
|
|
256
|
-
<summary><strong>Does it work with monorepos?</strong></summary>
|
|
257
|
-
|
|
258
|
-
Yes. Run `caliber init` from any directory. `caliber refresh` can update configs across multiple repos when run from a parent directory.
|
|
259
|
-
|
|
260
|
-
</details>
|
|
261
|
-
|
|
262
|
-
<details>
|
|
263
|
-
<summary><strong>Does it send my code anywhere?</strong></summary>
|
|
264
|
-
|
|
265
|
-
Scoring is fully local. Generation sends your project fingerprint (not source code) to whatever LLM provider you configure — the same provider your AI editor already uses. Anonymous usage analytics (no code, no file contents) can be disabled via `caliber config`.
|
|
266
|
-
|
|
267
|
-
</details>
|
|
268
|
-
|
|
269
|
-
## Add a Caliber badge to your repo
|
|
270
|
-
|
|
271
|
-
After scoring your project, add a badge to your README:
|
|
272
|
-
|
|
273
|
-
|
|
274
|
-
|
|
275
|
-
Copy this markdown and replace `94` with your actual score:
|
|
276
|
-
|
|
277
|
-
```
|
|
278
|
-
|
|
279
|
-
```
|
|
280
|
-
|
|
281
|
-
Color guide: `brightgreen` (90+), `green` (70-89), `yellow` (40-69), `red` (<40).
|
|
282
|
-
|
|
283
|
-
## LLM Providers
|
|
180
|
+
## 🔌 LLM Providers
|
|
284
181
|
|
|
285
182
|
No API key? No problem. Caliber works with your existing AI tool subscription:
|
|
286
183
|
|
|
@@ -337,12 +234,12 @@ export GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
|
|
|
337
234
|
|
|
338
235
|
</details>
|
|
339
236
|
|
|
340
|
-
## Requirements
|
|
237
|
+
## 📋 Requirements
|
|
341
238
|
|
|
342
239
|
- **Node.js** >= 20
|
|
343
240
|
- **One LLM provider:** your **Claude Code** or **Cursor** subscription (no API key), or an API key for Anthropic / OpenAI / Vertex AI
|
|
344
241
|
|
|
345
|
-
## Contributing
|
|
242
|
+
## 🤝 Contributing
|
|
346
243
|
|
|
347
244
|
See [CONTRIBUTING.md](./CONTRIBUTING.md) for detailed guidelines.
|
|
348
245
|
|
|
@@ -357,6 +254,6 @@ npm run build # Compile
|
|
|
357
254
|
|
|
358
255
|
Uses [conventional commits](https://www.conventionalcommits.org/) — `feat:` for features, `fix:` for bug fixes.
|
|
359
256
|
|
|
360
|
-
## License
|
|
257
|
+
## 📄 License
|
|
361
258
|
|
|
362
259
|
MIT
|
package/package.json
CHANGED