Shelem Scorekeeping MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SHELEM_MCP_LOG_LEVEL | No | Sets the level of diagnostics written to stderr. Same as the --log-level option. One of DEBUG, INFO, WARNING, ERROR, CRITICAL. | WARNING |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
Scores Shelem (شلم) hands. The server is stateless: keep the running game state yourself and pass the totals from before each hand.
Call calculate_shelem_score once per completed hand. Never compute a score yourself and never announce a winner the tool did not return.
Expand spoken abbreviated bids before calling: "20" means 120, "40" means 140. Only 100, 105, ..., 165 are legal bids.
Mak (مک) means the declarer captured exactly the bid; send outcome="mak".
A bid of 165 needs contract "shelem", "sare-shelem" or "normal".
If the tool returns an error, ask the user for the missing or contradictory fact instead of changing the numbers.
To correct an earlier hand, re-score every hand from that point on.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| calculate_shelem_scoreA | Score one completed Shelem hand and return both teams' new totals. Pass the cumulative scores from before the hand, the declaring team, the bid, and how many points were captured (defending_points, declarer_points, or outcome "mak"/"sweep"). The engine applies the doubled sweep award, failed and heavily failed bids, Shelem/sar-shelem (+330/-330/-660), the 1000-point rule for a defending team, and the 1165 winning threshold. All arithmetic is integer and deterministic. Invalid or contradictory input returns an error and no score. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 1 tool
There is only one tool, so there is no possibility of misselection. Its purpose (score a single completed Shelem hand) is unmistakable.
The single name calculate_shelem_score follows a clear snake_case verb_noun pattern with a domain qualifier. Nothing inconsistent to compare against.
A deterministic scoring engine can legitimately be a single atomic operation, and this tool is well-scoped and earns its place. However, one tool is borderline thin for a game domain that also involves game/round state, so it sits at the low end of the appropriate range.
The scoring surface itself looks thorough: mak/sweep doubling, failed and heavily failed bids, shelem/sar-shelem awards, the 1000-point rule, and the win threshold are all handled. Gaps are peripheral rather than core — there is no multi-hand or running-match aggregation and no way to inspect or explain the rules.