Cross-ecosystem agent behavioral telemetry SDK — report and read trust scores from the Dominion Observatory. The only MCP scoring network that accepts agent-reported runtime data.
Project description
dominion-observatory-sdk (Python)
One-line agent behavioral telemetry for MCP servers, LangChain tools, AutoGen agents, and CrewAI workers. Feeds the Dominion Observatory, the only MCP scoring network that accepts agent-reported runtime data instead of static GitHub/registry signals.
Install
Single file, stdlib only (Python 3.9+). No pip, no package manager:
curl -O https://sdk-cdn.sgdata.workers.dev/v1/observatory.py
Then drop observatory.py into your project and import from it directly.
A PyPI mirror will follow once the publish token is provisioned. The CDN is the canonical install channel today and every fetch fires anonymized adoption telemetry back into the Observatory.
Usage
from observatory import report, check_trust, instrument
SERVER_URL = "https://my-mcp-server.example.com/mcp"
# 1. Fire-and-forget telemetry in a tool handler
report(
server_url=SERVER_URL,
success=True,
latency_ms=142,
tool_name="get_holidays",
)
# 2. Convenience wrapper: measures latency automatically
result = instrument(SERVER_URL, "get_holidays", lambda: do_work())
# 3. Read a trust score before delegating to another server
score = check_trust("https://some-other-mcp.example.com/mcp")
if score.found and (score.trust_score or 0) >= 70:
call_that_server()
What gets sent
Exactly these five fields, and nothing else:
| Field | Example |
|---|---|
server_url |
https://my-mcp-server.example.com/mcp |
success |
true |
latency_ms |
142 |
tool_name |
get_holidays |
http_status |
200 |
Not sent: query arguments, tool output, user IDs, IP addresses, auth tokens, or any other payload. This satisfies Singapore PDPA and IMDA's agentic AI transparency requirements, and is compatible with EU AI Act Article 12 logging.
Why this exists
5+ MCP scoring platforms already exist (Glama, Smithery, MCP Scorecard, Nerq, Zarq), but all of them score servers from the outside using GitHub and registry metadata. None of them observe what happens when an agent actually calls a tool. The Dominion Observatory is the only network that collects agent-reported, cross-ecosystem runtime behavior — the only data that ages into compounding value because it can't be backfilled.
Every install of this SDK is a permanent telemetry pipeline.
Observatory API
| Purpose | Endpoint |
|---|---|
| Report interaction | POST https://dominion-observatory.sgdata.workers.dev/mcp (MCP JSON-RPC) |
| Trust score | GET https://dominion-observatory.sgdata.workers.dev/api/trust?url=... |
| Global stats | GET https://dominion-observatory.sgdata.workers.dev/api/stats |
License
MIT
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file dominion_observatory_sdk-0.1.0.tar.gz.
File metadata
- Download URL: dominion_observatory_sdk-0.1.0.tar.gz
- Upload date:
- Size: 4.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
8c23257a8396d6f4c2505cee5a8817b6e9a914330a7ee6483a06dd97c5855fc3
|
|
| MD5 |
87f48c7d59a8608bbdaa576319725b46
|
|
| BLAKE2b-256 |
0d5c2192ca9141132053b369c14a24f2148acff5d52d124e6535656bd4c19c31
|
File details
Details for the file dominion_observatory_sdk-0.1.0-py3-none-any.whl.
File metadata
- Download URL: dominion_observatory_sdk-0.1.0-py3-none-any.whl
- Upload date:
- Size: 4.6 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
549d457b9e781b0de578039325f5ffc5263f0f30da2d701f97459f5581d639c7
|
|
| MD5 |
9cc9acfbfa49d71105c181c408bf8d87
|
|
| BLAKE2b-256 |
ca963bdc9d809e214f7ca24bf710733619bf61f528c063111e7e6e983a88178d
|