agentsclimarketplace
Sign in

Si live

Skill bfowle/spirit-island/skills/si-live

Launch + interact with the interactive Spirit Island board-state tool. State lives at data/current-game.json and is shared between the web UI, this skill, and (Phase B onward) the MCP server.From its SKILL.md

Install
npx -y skills add bfowle/spirit-island --skill si-live

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.3 KB, 766 tokens by cl100k_base, as published. Nobody here has run it

si-live

Interactive Spirit Island board-state tool. Eliminates typed shorthand from the table workflow: the web UI writes data/current-game.json; skills (si-at-the-table, si-post-game) read the same file.

Prerequisites

  • Rust toolchain (for backend + MCP server).
  • Node 18+ / npm (for Vue 3 frontend).
  • One-time frontend install: cd tools/si-live/frontend && npm install.

How to launch (manual)

The server needs to run outside the Claude Code process — Claude's skills are stdin/stdout and can't host a long-lived web server. Brett launches it once per session:

./tools/si-live/start.sh

Opens two processes:

  • axum backend on http://127.0.0.1:7777 — owns data/current-game.json writes.
  • Vite dev server on http://127.0.0.1:5173 — proxies /api to the backend.

Brett opens the frontend URL in a browser, sets up the game (adversary, spirits, board), and clicks through state updates. The backend persists to data/current-game.json debounced.

How Claude uses this skill

Phase A (current — file read): every skill that needs live board state (si-at-the-table, si-post-game) reads data/current-game.json directly. Schema is documented at tools/si-live/common/src/schema.rs.

Phase B (planned — MCP): the si-mcp binary exposes get_game_state, compute_innate_threshold, compute_draw_probability, etc. as MCP tools. Skills switch to MCP calls instead of file reads. File-read remains the fallback.

When to invoke

  • User is about to play a game and asks Claude to help track / analyze in real time.
  • User says "launch the tool" / "start si-live" / "open the board".
  • si-at-the-table detects the state file is stale or missing and wants to prompt launch.

Do not auto-launch on every Claude invocation — Brett runs the tool when he wants it up. If the state file is missing or stale, tell Brett to run ./tools/si-live/start.sh; don't try to spawn it yourself.

State schema summary

See tools/si-live/common/src/schema.rs for the authoritative types. Key shapes:

  • GameState: { version, round, phase, setup, pools, spirits, board_state, log }
  • Spirit: { energy, card_plays, presence_on_board, presence_on_track_energy, presence_on_track_cardplay, elements_this_turn, hand, discard, forgotten, played_this_turn, growth_options }
  • Board: { lands: { "<id>": Land } }
  • Land: { terrain, coastal, explorers, towns, cities, dahan, blight, tokens }

Tracks are represented as remaining-covered-slot arrays. The head of the array is the next slot that will be revealed. Starting Energy/CP come from presence_energy_track[0] / presence_cardplay_track[0] — this is the invariant si-rules-check relies on.

Related skills

  • si-rules-check — validate proposed plays against starting state + growth options.
  • si-at-the-table — read the live state + return Rei-format option trade analysis.
  • si-post-game — pull end-state from data/current-game.json and append to playlog.
  • si-wiki-fetch — authoritative source for spirit/card/adversary data.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most mcp tooling skills give in 766 tokens

Counted across 638 of the 750 authors here whose files we hold, read 2026-08-07

  • Create ten complex or independent read-only evaluation questionsin 69 of 638, across 15 files
  • Test servers using MCP Inspectorin 61 of 638, across 19 files
  • Provide actionable error messages with specific next stepsin 54 of 638, across 12 files
  • Prioritize comprehensive API coverage over specific workflows or workflow toolsin 54 of 638, across 12 files
  • Use TypeScript and Streamable HTTP for remote servers or clientsin 54 of 638, across 8 files
  • Define structured output schemas where possiblein 50 of 638, across 8 files
  • Use Zod or Pydantic for input schemasin 47 of 638, across 5 files
  • Fetch MCP specification pages with markdown suffixin 46 of 638, across 4 files
  • Load framework documentation using WebFetchin 45 of 638, across 3 files
  • Verify each evaluation answer independentlyin 45 of 638, across 3 files
  • Implement API client with authentication and paginationin 45 of 638, across 3 files
  • Define input schemas with validationin 27 of 638, across 9 files

Said here and by no other author read

  • Read data/current-game.json for live board state
  • Consult schema.rs for authoritative state types
  • Read presence track head for starting energy or cardplay
  • Prompt Brett to run start.sh if state file is missing or stale

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,851. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.