@the-17/agentsecrets 1.1.0 → 1.1.2
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 +113 -0
- package/package.json +1 -1
package/README.md
ADDED
|
@@ -0,0 +1,113 @@
|
|
|
1
|
+
# AgentSecrets
|
|
2
|
+
|
|
3
|
+
> **Zero-knowledge secrets infrastructure built for AI agents to operate, not just consume.**
|
|
4
|
+
|
|
5
|
+
Every other secrets tool was built for humans to provision credentials to agents. AgentSecrets was built for agents to manage credentials themselves — without ever seeing a single value.
|
|
6
|
+
|
|
7
|
+
[](https://opensource.org/licenses/MIT)
|
|
8
|
+
[](https://go.dev/)
|
|
9
|
+
[](https://clawhub.ai/SteppaCodes/agentsecrets)
|
|
10
|
+
|
|
11
|
+
**[Official Website](https://agentsecrets.theseventeen.co)** | **[Engineering Blog Series](https://engineering.theseventeen.co/series/building-agentsecrets)**
|
|
12
|
+
|
|
13
|
+
---
|
|
14
|
+
|
|
15
|
+
## What This Is
|
|
16
|
+
|
|
17
|
+
Most secrets tools treat AI agents as consumers, something that receives a credential and uses it. AgentSecrets treats the agent as an operator.
|
|
18
|
+
|
|
19
|
+
Your agent checks its own status, notices a secret is out of sync, pulls the latest from the cloud, makes the authenticated API call, and audits what it did. All of this without ever knowing a single credential value.
|
|
20
|
+
|
|
21
|
+
```bash
|
|
22
|
+
# An AI agent managing its own secrets workflow autonomously
|
|
23
|
+
|
|
24
|
+
agentsecrets status # what workspace, project, last sync?
|
|
25
|
+
agentsecrets secrets diff # anything out of sync?
|
|
26
|
+
agentsecrets secrets pull # sync from cloud to keychain
|
|
27
|
+
agentsecrets secrets list # what keys are available?
|
|
28
|
+
agentsecrets call \
|
|
29
|
+
--url https://api.stripe.com/v1/balance \
|
|
30
|
+
--bearer STRIPE_KEY # make the authenticated call
|
|
31
|
+
agentsecrets proxy logs # audit what just happened
|
|
32
|
+
```
|
|
33
|
+
|
|
34
|
+
The agent ran the entire credentials workflow. It never saw `sk_live_51H...`. Not at any step.
|
|
35
|
+
|
|
36
|
+
This is what it means to be built for the agentic era — not bolted onto it.
|
|
37
|
+
|
|
38
|
+
---
|
|
39
|
+
|
|
40
|
+
## Installation
|
|
41
|
+
|
|
42
|
+
**Homebrew (macOS / Linux):**
|
|
43
|
+
```bash
|
|
44
|
+
brew install The-17/tap/agentsecrets
|
|
45
|
+
```
|
|
46
|
+
|
|
47
|
+
**npm / npx:**
|
|
48
|
+
```bash
|
|
49
|
+
npm install -g @the-17/agentsecrets
|
|
50
|
+
# or without installing
|
|
51
|
+
npx @the-17/agentsecrets init
|
|
52
|
+
```
|
|
53
|
+
|
|
54
|
+
**pip:**
|
|
55
|
+
```bash
|
|
56
|
+
pip install agentsecrets-cli
|
|
57
|
+
```
|
|
58
|
+
|
|
59
|
+
**Go:**
|
|
60
|
+
```bash
|
|
61
|
+
go install github.com/The-17/agentsecrets/cmd/agentsecrets@latest
|
|
62
|
+
```
|
|
63
|
+
|
|
64
|
+
---
|
|
65
|
+
|
|
66
|
+
## Quick Start
|
|
67
|
+
|
|
68
|
+
```bash
|
|
69
|
+
# Set up your account (first time) or initialise a new project (returning user)
|
|
70
|
+
agentsecrets init
|
|
71
|
+
|
|
72
|
+
# Create a project
|
|
73
|
+
agentsecrets project create my-app
|
|
74
|
+
|
|
75
|
+
# Store credentials — values go to OS keychain, never to disk
|
|
76
|
+
agentsecrets secrets set STRIPE_KEY=sk_live_51H...
|
|
77
|
+
agentsecrets secrets set OPENAI_KEY=sk-proj-...
|
|
78
|
+
agentsecrets secrets set DATABASE_URL=postgresql://...
|
|
79
|
+
|
|
80
|
+
# Or push your existing .env all at once
|
|
81
|
+
agentsecrets secrets push
|
|
82
|
+
|
|
83
|
+
# Authorize the domains your agents can reach
|
|
84
|
+
agentsecrets workspace allowlist add api.stripe.com api.openai.com
|
|
85
|
+
|
|
86
|
+
# Connect your AI tool
|
|
87
|
+
npx @the-17/agentsecrets mcp install # Claude Desktop + Cursor
|
|
88
|
+
agentsecrets proxy start # Any agent via HTTP
|
|
89
|
+
openclaw skill install agentsecrets # OpenClaw
|
|
90
|
+
|
|
91
|
+
# Or inject secrets as env vars into any process
|
|
92
|
+
agentsecrets env -- stripe mcp
|
|
93
|
+
agentsecrets env -- node server.js
|
|
94
|
+
agentsecrets env -- npm run dev
|
|
95
|
+
```
|
|
96
|
+
|
|
97
|
+
Your agent now has full API access. It will never see a credential value.
|
|
98
|
+
|
|
99
|
+
---
|
|
100
|
+
|
|
101
|
+
## Full Command Reference
|
|
102
|
+
|
|
103
|
+
See the [Official Documentation](https://agentsecrets.theseventeen.co) for the full command reference and architecture deep dives.
|
|
104
|
+
|
|
105
|
+
---
|
|
106
|
+
|
|
107
|
+
## License
|
|
108
|
+
|
|
109
|
+
MIT — see [LICENSE](https://github.com/The-17/agentsecrets/blob/main/LICENSE)
|
|
110
|
+
|
|
111
|
+
Built by [The Seventeen](https://theseventeen.co)
|
|
112
|
+
|
|
113
|
+
**The agent operates it. The agent never sees it.** ⭐
|