CliDeck MCP server

• Registry namecom.clideck/network-knowledgeverified in the MCP registry

The registry name sits under clideck.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-10-24
• Tools20
• Server calls itselfCliDeck MCP — Network Knowledge
• Server version0.8.5
• Protocol version2025-06-18

On 2026-10-24 the server answered initialize and tools/list without a sign-in. First answer in 216 ms.

• Its tools do none of the product's use cases we checked: provision and recover network devices with local services; log each device run against asset tags and serial numbers.
• Checked on 4 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 4 so far
• What an agent can do with it0 of 2 use cases
• Use3 GitHub stars
• Answeredmeasuring4 of the 7 daily checks it takes to print a share
• Median first answer257 ms
• Space in an agent's context≈ 3,748 tokensthe definitions of all its tools (20), which an agent carries on every turn
• Tool definitionsB · 84 of 100
• Typed results20 of 20 toolsdeclared outputSchema: results come typed, read once

• Parameters without a description: in 19 of 20

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• report_knowledge_gap

  Record the original network question for priority research. Provide missing context or explain why an answer was not useful. Feedback alone does not change or quarantine published knowledge.

• report_domain_knowledge_gap

  Retain the original technical question, independent OS/component versions and missing details for priority learning. Does not alter published revisions.

• get_learning_status

  Check a learning ID returned by a knowledge query or report_knowledge_gap. Published means the original question has a complete answer; repeat the original query to read it.

• list_knowledge_domains

  List installed, compatible knowledge domain packs and their declared capabilities.

• describe_knowledge_domain

  Return the context and public-record contracts for one installed knowledge domain.

• query_domain_knowledge

  Run deterministic search in one domain using that pack’s validated context and public-record schema. Network queries may create or update a tracked research demand for missing or incomplete knowledge.

• resolve_network_context

  Resolve vendor, model/platform, portable software family, operating system, and version context.

• query_network_knowledge

  Return deterministic, version-scoped commands, diagnostics, and concepts for a network question. Missing or incomplete knowledge may create or update a tracked research demand. No device commands are executed.

• get_knowledge_provenance

  Return safe source metadata for up to five active knowledge revisions.

• get_network_workflow

  Return a deterministic, ordered workflow including safety, verification, and rollback. Missing or incomplete knowledge may create or update a tracked research demand. No device commands are executed.

• request_expert_answer

  Create a durable research task when deterministic knowledge has no applicable answer.

• get_expert_task

  Poll an expert research task by opaque ID and access token.

• continue_expert_task

  Provide bounded additional input requested by the researcher.

• cancel_expert_task

  Cancel a non-terminal expert task.

• submit_feedback

  Submit correctness, freshness, safety, or completeness feedback without including private logs.

• analyze_device_snapshot

  Detect vendor, model, operating system, and version from bounded CLI output while redacting sensitive values. Input is not stored.

• review_network_change

  Deterministic advisory review of commands or a configuration diff. It always returns available guidance and never executes a command. Missing or incomplete knowledge may create or update a tracked research demand.

• verify_network_change

  Evaluate bounded before/after outputs against the signed verification plan returned by review_network_change.

• advise_network_upgrade

  Return exact-model, exact-version upgrade prerequisites, risks, checks, and rollback without downloading software. Missing or incomplete knowledge may create or update a tracked research demand. No device commands are executed.

• analyze_network_path

  Build a bounded topology and packet-path graph from supplied CDP, LLDP, route, and traceroute outputs without storing them.