@battlegrid/mcp-server 31.2.35 → 31.2.37

package/README.md

@@ -24,6 +24,49 @@ Seeing package `31.x` alongside handshake `battlegrid@33.x` — the package **be
 
 **What this changes for you:** nothing about how you call anything. Upgrading the package no longer waits on a server deploy, and a server deploy no longer strands you on a package that names the wrong contract — reconnect and the announcement follows. **Contract breaking-change notes are no longer keyed to package versions**, since a contract move is no longer a release here; the v11-and-earlier notes below are kept as history, and the live vocabulary is always discovery.
 
+## Contract history — v63.2
+
+Purely additive again, and smaller: **one optional input field.** Drafts now cover a strategy that
+does not exist yet, so a playbook you author in chat survives between compiles instead of starting
+over each time.
+
+### Wider input — one optional field on one tool
+
+- **`compile_strategy_plan`'s CREATE arm gains an optional `strategyId`**, naming the create draft
+  this compile continues. Omit it and the server mints an id exactly as it always has, so nothing
+  you send today breaks.
+
+- **Send it to accumulate; omit it to start fresh.** A first CREATE names nothing and mints; stage
+  that plan and the draft opens at the minted id, which `list_strategy_drafts` and the plan itself
+  both report. Name that id on the next CREATE and the second plan carries the same identity, so the
+  one draft accumulates rather than a second opening beside it.
+
+- **WHY YOU HAVE TO SAY IT.** The server will not choose for you. A compile knows only who you are —
+  it has no conversation id — and you may hold several create drafts, so picking one could write this
+  playbook's values over another's. A compile that names none mints a third id rather than guessing.
+
+- **It names a draft; it does not choose an id.** It is accepted only as the identity of a create
+  draft you own. An id naming no such draft is refused, as is one naming a draft for a strategy that
+  already exists.
+
+### Behaviour changes behind unchanged schemas
+
+- **`stage_strategy_plan` admits a CREATE plan**, opening the create draft at the plan's own id when
+  none exists yet. Its input is still `{ planToken }` alone, and the contested-axis refusal applies
+  identically.
+
+- **`list_strategy_drafts` rows gain `baseRevision`.** It is `null` for a create draft — a strategy
+  that does not exist has no committed revision to be based on — and that null is the only thing that
+  distinguishes one. There is no `kind`, `isNew` or `status` field beside it.
+
+- **A create draft reserves nothing.** It consumes no quota and holds no name, so both are decided
+  when the strategy is actually created. Holding drafts past your limit is legal; creating past it is
+  not.
+
+### Vocabulary
+
+`axes.create` gains `strategyId`. No domain gains a value, and `toolCount` stays 122.
+
 ## Contract history — v63.1
 
 Purely additive: **four new tools, nothing you send today changes.** A strategy now has an

package/dist/index.d.ts

@@ -49,7 +49,7 @@ import { type Implementation, type Prompt, type Resource } from '@modelcontextpr
  * being asked. Move it for a change to THIS package — a proxy fix, a dependency bump, a docs
  * correction. Never move it to track the server.
  */
-export declare const PACKAGE_VERSION = "31.2.35";
+export declare const PACKAGE_VERSION = "31.2.37";
 export declare const DEFAULT_URL = "https://mcp.battlegrid.trade/mcp";
 export interface EnvConfig {
     apiKeys: string[];

package/dist/index.js

@@ -52,7 +52,7 @@ import { ListToolsRequestSchema, CallToolRequestSchema, ListPromptsRequestSchema
  * being asked. Move it for a change to THIS package — a proxy fix, a dependency bump, a docs
  * correction. Never move it to track the server.
  */
-export const PACKAGE_VERSION = '31.2.35';
+export const PACKAGE_VERSION = '31.2.37';
 export const DEFAULT_URL = 'https://mcp.battlegrid.trade/mcp';
 const MAX_RETRIES = 3;
 const RETRY_DELAYS_MS = [2000, 4000, 8000];

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@battlegrid/mcp-server",
-  "version": "31.2.35",
-  "description": "BattleGrid MCP server \u2014 play crypto prediction games from AI agents",
+  "version": "31.2.37",
+  "description": "BattleGrid MCP server — play crypto prediction games from AI agents",
   "type": "module",
   "main": "dist/index.js",
   "bin": {

package/skills/EXPORT.json

@@ -1,12 +1,12 @@
 {
   "generator": "battlegrid-app/server/scripts/export-mcp-skills.mjs",
-  "contractVersion": "63.0.0",
+  "contractVersion": "63.2.0",
   "files": {
     "battlegrid-agent-management/SKILL.md": "d56834fa2016ec9371994999fa13a3dbb69906efc54db522e10fd2eb0ebb6044",
     "battlegrid-arena-play/SKILL.md": "4cfaa92a8d0dcaf477b227abd774e7aa8642c185d72108dfd401be0cf8aecd7f",
     "battlegrid-market-analysis/SKILL.md": "22fde9c4eac0261c89a8056ab8b20b66fcbe1fd58447fa0ac84fd410b4d134d7",
     "battlegrid-radar-deployment/SKILL.md": "d2a012b154e1e529db19529bd7c3d75d069f4e410d5804add51165f0f0902b2e",
-    "battlegrid-strategy-authoring/SKILL.md": "66ed78d8e83401586b6d0ec33ce59b0fdc8b22b04a3db0ab529cc77779bfe1c9",
+    "battlegrid-strategy-authoring/SKILL.md": "2ad66540d406fdf447118396821aa2aeb3510045a79d397846b83c7c797f33a1",
     "battlegrid-strategy-doctor/SKILL.md": "d0b9e9e07648246d5a23b0c17183278b70b5bcbb5f472ac7da13d03e42a6c6c6",
     "battlegrid-strategy-examples/SKILL.md": "17e8cf8040813622bd5b82165c7af82e03287fee1f6aac974e0b6c584580dd27",
     "battlegrid-strategy-examples/references/playbooks.md": "35c543c090d00ba19e4bcf0a80502afac7a4954643fc535de0b24381c91ecbf4",

package/skills/battlegrid-strategy-authoring/SKILL.md

@@ -52,6 +52,18 @@ conversation has changed it. A re-read returns the same bytes, and both copies t
 later step, so the player pays for the same payload twice and keeps paying for it. If you need a
 detail you did not keep, scroll back rather than re-fetching.
 
+**Drafts are carved out of that rule.** A draft is the player's unsaved work, and another surface —
+their builder, another device, an earlier plan you staged — can change it while you work. Read it
+again whenever you are about to act on it.
+
+**Start from what they are already part-way through.**
+
+- Holding no strategy id? Call `list_strategy_drafts` first. If a draft comes back, say what it is,
+  when it was last touched and which surface touched it, and offer to continue it — never start a
+  second edit beside one the player has open.
+- Holding an id? Call `get_strategy_draft` before you propose anything. A draft means the player is
+  mid-edit.
+
 ### 2. Lock the spec before you build anything
 
 **First, check the ask is expressible at all — and note that you cannot know until you have looked.**
@@ -187,9 +199,30 @@ returned, as numbers, not buried in prose.
 There is no backtest here and no expected-frequency figure. Do not imply one. What you have is a
 point-in-time reading, and you say so.
 
-### 6. Confirm, then apply
+### 6. Stage or apply
+
+**A draft decides which of the two you are doing.**
+
+- **The player holds a draft for this strategy** → call `stage_strategy_plan` with the compile's
+  `planToken` and nothing else. The plan's own changed values land in their draft as proposed
+  changes, and *their* save is the consent. Committing instead would end a session they are in the
+  middle of. Staging does not spend the plan: the same token still applies while it lives.
+- **They hold none** → compile → confirm → apply, as below.
+
+When you compile over a draft, the diff names **every** axis the plan would commit, including
+unsaved work the draft already carried before your compile. Read that list out in the
+confirmation as the plan's own: the player is approving all of it, not only what you proposed.
+
+If a staging call is refused as contested, it names the axes the player's own hand changed after
+your compile. Do not retry it and do not work around it — compile again so the new plan absorbs
+those edits, then stage that one.
+
+**Offer to discard a draft only on the player's explicit word**, never on your own judgement that
+it looks stale, and never batched into another act. State when it was last touched and which
+surface touched it, ask, and call `discard_strategy_draft` with `confirm: true` only if they say
+so. The unsaved values are gone afterwards and there is no other copy.
 
-**In a conversation the Strategy Builder hosts, this section does not apply: compile and stop.**
+**In a conversation the Strategy Builder hosts, the apply half does not apply: compile and stop.**
 The compiled plan lands on the player's rail as unsaved changes and their Review & save is the one
 consent. Report that it is staged, name the axes, and present no apply confirmation — the server
 refuses `apply_strategy_plan` from such a conversation as a tool error, so calling it spends an op