wordle-solver
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| WORDLE_API_BASE | No | Point at a different deployment of the solver |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
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
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_game_statsA | Get the state of a Wordle game before any guess has been made. Call this at the START of a game, before the first guess: it returns the size of the guess dictionary, how many words are eligible answers, and the starting uncertainty in bits. Do not call recommend_guesses with an empty guess history - scoring the full dictionary is too expensive, so open with a standard high-information first word instead. |
| recommend_guessesA | Get ranked next-guess recommendations for an in-progress Wordle game. Call this AFTER each guess, passing the full game so far: every guessed word and its color feedback, in order. Feedback is one string per guess, one character per letter: 'g' green (right letter, right spot), 'y' yellow (in the word, wrong spot), 'x' gray (not in the word). Example: the guess CRANE showing gray-yellow-gray-green-gray is words=["crane"], feedback=["xyxgx"]. Returns the top recommendations ranked by expected information gain (lower bits = better guess), the best guesses that are also eligible answers, and game-state statistics including remaining possibilities and uncertainty in bits. Prefer a word from best_answers when few answers remain; prefer best_overall when many remain and you want to maximize information. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| methodology | How the solver ranks guesses - read before explaining recommendations. |
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: get_game_stats initializes a game, while recommend_guesses drives in-progress gameplay. Each description explicitly states when to call the tool and what it must not be called with, leaving no ambiguity about which to invoke at any stage.
Both tools follow a verb_noun pattern (get_game_stats, recommend_guesses). The verbs differ (get vs recommend) but this reflects distinct actions rather than inconsistency; this is a minor stylistic variation within an otherwise consistent scheme.
Two tools is on the thin side for a solver server. The set covers game initialization and guidance, but a solver could reasonably also expose a dedicated 'get opening word' tool or a 'filter results' capability. That said, two focused tools for a tightly scoped purpose is defensible.
The surface covers the start-state and the recommendation loop, which are the core needs. However, get_game_stats doesn't itself return a first guess (it explicitly instructs against calling recommend_guesses with empty history), leaving the opening move unexplained. A tool to provide the standard high-information first word would close the gap between game start and first guess.