Read-only Exoscale advisor MCP server — docs search plus list-only live catalogue queries, structurally incapable of mutation.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"exoscale-mcp-advisor","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# exoscale-mcp-advisor A **read-only** [Model Context Protocol](https://modelcontextprotocol.io) server that lets an MCP-capable agent learn about Exoscale: search the verified connector documentation and run **list-only** live catalogue queries (zones, instance types, templates, managed-database plans, SKS versions). It is, by construction, incapable of mutating any cloud resource. > **Status: released.** Eight read-only tools, the stdio server, and the > four-layer test suite (structural no-mutation, mocked-connector, > protocol-level, gated live smoke) are in place and green; published to PyPI as > [`exoscale-mcp-advisor`](https://pypi.org/project/exoscale-mcp-advisor/). The > full design is in [`docs/mcp-advisor-design.md`](docs/mcp-advisor-design.md); > release history is in [`CHANGELOG.md`](CHANGELOG.md). It builds on [`exoscale-connector`](https://github.com/ralle-lang/exoscale-python-connector): the knowledge it serves is read from that package's bundled reference, and the live queries reuse the connector's read-only `list` clients. Nothing about Exoscale is hardcoded or duplicated here. --- ## General description The advisor targets the learning path, not the execution path. An agent can ask "what instance types exist in `at-vie-1` right now?" or "how do I create a security group with the connector?" and get live data plus verified docs, while the server stays structurally unable to create, change, or delete anything. Infrastructure changes stay the human's job, performed with reviewed, idempotent code. Tool surface (see design §3): | Tool | Credentials | Purpose | |------|:-----------:|---------| | `search_docs(query)` | — | Ranked sections from the connector's reference bundle. | | `get_asset_page(asset_type)` | — | Full reference page for one asset type. | | `list_asset_types()` | — | The asset-type index (slug + heading) for discovery. | | `list_zones()` | ✔ | Live list of zones. | | `list_instance_types(zone)` | ✔ | Live list of instance types…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.