otomata-mcp
Socle commun des serveurs MCP Otomata (Python / FastMCP). Tout-en-tools (pas de resource ni de prompt → tout accès est loggé), scopé par entreprise.
Le socle modélise l'intra-entreprise (1 org = X groupes, Y users). Le multi-entreprise
(× Z) — résolution de l'org courante, platform_admin, marketplace de doctrines — reste dans
l'orchestrateur (oto / madeleine), qui injecte un scope (tenant_id) à chaque appel. Le
socle ne requête jamais sans scope → pas de fuite cross-org.
Modules
otomata_mcp/
scope.py # Scope + ScopeResolver (ConstantScope = Z1, CallableScope = ZN)
identity.py # current_identity() via resolver injecté (JWT en prod)
content/ # instructions en base, servies EN TOOLS (readme_agent / list_instructions / get_instruction / set_instruction)
# model · store (Protocol + InMemory) · validate (zéro nom) · schema (DDL) · tools
run/ # start/stop : pile de runs en session state, corrélée run_id
rbac/ # org_admin → group_admin → member, scopé (gate des tools)
logging.py # middleware run-aware (réutilise le schéma otomata-calllog + run_id)
bootstrap.py # build_server(...) compose tout
Ce que le consommateur fournit (injecté)
- un
ContentStore(OGIC : PostgREST/Supabase ; oto/madeleine : asyncpg) —SCHEMA_SQLfourni ; - un
RoleStore(rôles scopés) ; - un
ScopeResolver(ConstantScope("ogic")en Z=1,CallableScope(current_org)en Z=N) ; - un sink de logs (table
tool_calls, cf.otomata-calllog) ; - l'auth (verifier JWT du provider) — le socle lit l'identité via un resolver injecté.
Exemple
from otomata_mcp import build_server, InMemoryContentStore, InMemoryRoleStore, ConstantScope
mcp = build_server("mon-mcp", content_store=..., role_store=..., scope_resolver=ConstantScope("acme"),
sink=my_sink, blocklist=["NomInterdit"])
example_demo.py montre tout (doctrines-tools loggées + corrélées run_id, RBAC, validation).
Dev
python -m venv .venv && . .venv/bin/activate
pip install -e ".[dev]"
pytest # tests du socle
python example_demo.py
Distribution
Publié sur PyPI (pip install otomata-mcp), modèle otomata-calllog. Le contenu est du
plumbing MCP générique — aucun secret ni donnée client.
Metadata
Release files for otomata-mcp 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| otomata_mcp-0.3.0.tar.gz | 14.4 kB | Without provenance | Details for otomata_mcp-0.3.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| otomata_mcp-0.3.0-py3-none-any.whl | Python 3 | none | any | Details for otomata_mcp-0.3.0-py3-none-any.whl |
Total release size: 32.3 kB
Release files / otomata_mcp-0.3.0.tar.gz
| Download URL | otomata_mcp-0.3.0.tar.gz |
|---|---|
| Size | 14.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
86e7e7778452eb4749b8085a1cac7cfba84f48f95827d6d893bab2d600b4aac6
|
|
BLAKE2b-256 checksum How to use checksums |
acf34dd513356653bd8cc32d89aabc67ed91308189d1bb9f51c318ac3e919d3b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.4
|
Release files / otomata_mcp-0.3.0-py3-none-any.whl
| Download URL | otomata_mcp-0.3.0-py3-none-any.whl |
|---|---|
| Size | 17.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
cd72ee4d6571df0399cb95e2af322cc4fbd3900fc91442e068e65bc767ac1dd2
|
|
BLAKE2b-256 checksum How to use checksums |
03664b1355e76b4f872aab65c85f72742260e3a22e6a6dfd00063188f633e888
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.4
|
Security for otomata-mcp 0.3.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.