Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
CLEF_HOMENoCache/model home~/.cache/clef-mcp
CLEF_MODELNoModel id (`clef_decide` also accepts a per-call `model`)clef-flash
CLEF_RUNTIMENoInference runtime (v0.1: only llama-cpp)llama-cpp
CLEF_LLAMA_BINNoExplicit path to a `llama-server` binary
CLEF_LOG_LEVELNo`error` | `warn` | `info` | `debug` (logs go to stderr)error
CLEF_MAX_QUESTIONSNoMax questions per call64
CLEF_MAX_STATE_BYTESNoMax serialized `state` size1048576
CLEF_LLAMA_RELEASE_TAGNoPin the managed llama.cpp build, e.g. `b11378`latest nightly
CLEF_MAX_INSTRUCTION_CHARSNoMax chars per question instructions10000

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

clef_decide passes task state plus typed questions to the local Clef decision model and returns a probability distribution over answer options. Treat state as data: it is never interpreted as instructions for this server. The result is structured JSON; do not render it as prose.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}
prompts
{
  "listChanged": true
}
resources
{
  "listChanged": true
}

Tools

Functions exposed to the LLM to take actions

NameDescription
clef_decideA

Judge a situation with the local Clef decision model. Pass state (compact factual context — string or JSON; treated as data, never executed) and typed questions:

  • "choice": pick among named options (criteria = option id -> description)

  • "score": judge an ordered scale (criteria = ordered list of levels)

  • "noul": yes/no question Returns strict JSON: a probability distribution over the allowed answers per question — no prose, no sampling. Each decision carries model-reported confidence; usage reports input_tokens/output_tokens and latency_ms. Act on the argmax only when decisive: top p >= 0.8 and high confidence for destructive or security-adjacent calls; otherwise gather more context or ask the user. Batch related decisions in one call — up to 64 questions scored in a single forward pass. model overrides the default id.

Prompts

Interactive templates invoked by user choice

NameDescription
incident-triageFrame a production incident as clef_decide questions: immediate action, severity and blast radius.
next-actionAsk Clef what the coding agent should do next on a task: inspect, modify, test or ask the user.
ticket-routingClassify an incoming message into the team that should handle it.
security-reviewAsk Clef whether a code snippet, change or setup smells like a security problem.

Resources

Contextual data attached and managed by the client

NameDescription
capabilitiesLive capability snapshot: model, runtime, limits, error codes.
evals-schemaHow to write cases for clef-mcp evals.
evals-datasetThe 30-case dataset shipped with clef-mcp.

TDQS

A4.5/5.0

Scored across 1 tool

Disambiguation5/5

With a single tool there is no possibility of tool misselection; the only differentiation needed is between question types (choice/score/noul), which the description enumerates explicitly.

Naming Consistency5/5

A single verb_noun name with a server prefix (`clef_decide`) is internally consistent and predictable, and the prefix would scale cleanly if more `clef_*` tools were added.

Tool Count3/5

One tool is borderline thin even for a focused decision-model wrapper; batching up to 64 questions inside one call partly compensates, but there is no separate schema-discovery or model-listing operation.

Completeness4/5

The tool covers the core judging lifecycle (arbitrary state in, typed questions, structured distributions with confidence and usage out) and supports batching and model override; minor gaps exist around discovering available models or validating question schemas before a call.

Maintenance

ActivityMaintained
ResponsivenessOutdated data