@gala-chain/launchpad-mcp-server 1.2.13 → 1.2.15

package/CHANGELOG.md

@@ -1,13 +1,27 @@
 # Changelog
 
+## 1.2.15
+
+### Patch Changes
+
+- Fix memory leaks and enable incremental builds
+  - Fix WebSocket onAny listener cleanup
+  - Fix HTTP interceptor cleanup
+  - Enable incremental TypeScript builds (50% faster rebuilds)
+  - Add comprehensive performance documentation
+
+- Updated dependencies
+  - @gala-chain/launchpad-sdk@3.7.1
+
 All notable changes to the Gala Launchpad MCP Server will be documented in this file.
 
-The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
-and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project
+adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [1.2.4] - 2025-10-02
 
 ### Changed
+
 - **Internal code quality improvements** - Schema consolidation across all 37 MCP tools
   - Created centralized `common-schemas.ts` with 15+ reusable schema definitions
   - Updated all tools to use shared schemas (TOKEN_NAME_SCHEMA, PRIVATE_KEY_SCHEMA, etc.)
@@ -17,6 +31,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Updated SDK dependency from ^3.5.1 to ^3.5.2
 
 ### Benefits
+
 - ✅ Improved code maintainability and consistency
 - ✅ Single source of truth for all schema definitions
 - ✅ Easier to maintain and update validation rules
@@ -25,9 +40,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.2.3] - 2025-10-01
 
 ### Documentation
+
 - **Enhanced trading workflow documentation** - Critical updates to prevent transaction failures
   - Added new Gotcha #1: Missing reverseBondingCurveFee Parameter with real-world failure examples
-  - Updated trading workflows (buy & sell) to emphasize BOTH expectedAmount AND reverseBondingCurveFee requirements
+  - Updated trading workflows (buy & sell) to emphasize BOTH expectedAmount AND
+    reverseBondingCurveFee requirements
   - Updated tool descriptions for `gala_launchpad_buy_tokens` and `gala_launchpad_sell_tokens`
   - Changed `maxAcceptableReverseBondingCurveFee` description from "Optional" to "RECOMMENDED"
   - Added explicit warning: "Omitting this may cause transaction failures"
@@ -35,30 +52,39 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Fixed type parameter references from 'token' to 'exact' throughout documentation
 
 ### Benefits
+
 - ✅ AI agents will now include reverseBondingCurveFee parameter by default
 - ✅ Reduced transaction failures due to missing fee parameter
 - ✅ Clearer workflow examples with complete parameter extraction
 - ✅ Better troubleshooting guidance with real failure examples
 
 ### Context
-This update addresses a recurring issue where AI agents forgot to pass the `reverseBondingCurveFee` parameter, causing "FAILED" transaction statuses. The documentation now treats this parameter as critical rather than optional.
+
+This update addresses a recurring issue where AI agents forgot to pass the `reverseBondingCurveFee`
+parameter, causing "FAILED" transaction statuses. The documentation now treats this parameter as
+critical rather than optional.
 
 ## [1.2.2] - 2025-10-01
 
 ### Changed
+
 - **Tool parameter standardization** - Renamed `walletAddress` to `address` in 6 MCP tools
-  - Updated tools: `gala_launchpad_fetch_gala_balance`, `gala_launchpad_fetch_token_balance`, `gala_launchpad_fetch_profile`, `gala_launchpad_update_profile`, `gala_launchpad_upload_profile_image`
+  - Updated tools: `gala_launchpad_fetch_gala_balance`, `gala_launchpad_fetch_token_balance`,
+    `gala_launchpad_fetch_profile`, `gala_launchpad_update_profile`,
+    `gala_launchpad_upload_profile_image`
   - All tools now accept both `0x...` (Ethereum) and `eth|...` (GalaChain) formats
   - Address formats automatically normalized by SDK
   - Error messages updated to reference `address` field
 - Updated SDK dependency from ^3.5.0 to ^3.5.1
 
 ### Documentation
+
 - Updated AI-AGENT-PATTERNS.md with new `address` parameter examples
 - Updated validation patterns table with address auto-normalization info
 - Error handler now references `address` instead of `walletAddress`
 
 ### Benefits
+
 - ✅ AI agents can use natural `0x...` addresses without manual conversion
 - ✅ Consistent parameter naming across all tools
 - ✅ Better developer experience with automatic format handling
@@ -66,11 +92,14 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 ## [1.2.0] - 2025-10-01
 
 ### Added
-- **Multi-wallet support via privateKey override** - All signing operations now accept optional `privateKey` parameter
+
+- **Multi-wallet support via privateKey override** - All signing operations now accept optional
+  `privateKey` parameter
   - Enables AI agents to execute operations from different wallets without restarting the MCP server
   - Added `privateKey` parameter to 9 MCP tools:
     - Trading: `gala_launchpad_buy_tokens`, `gala_launchpad_sell_tokens`
-    - Token Creation: `gala_launchpad_launch_token`, `gala_launchpad_upload_token_image`, `gala_launchpad_upload_profile_image`
+    - Token Creation: `gala_launchpad_launch_token`, `gala_launchpad_upload_token_image`,
+      `gala_launchpad_upload_profile_image`
     - Transfers: `gala_launchpad_transfer_gala`, `gala_launchpad_transfer_token`
     - Social & Profile: `gala_launchpad_post_comment`, `gala_launchpad_update_profile`
   - `privateKey` must be in format: `'0x' + 64 hexadecimal characters`
@@ -78,9 +107,11 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
   - Powered by SDK v3.5.0 multi-wallet support
 
 ### Changed
+
 - Updated SDK dependency from ^3.3.4 to ^3.5.0
 
 ### Benefits
+
 - ✅ Test multi-wallet trading scenarios in AI agents
 - ✅ Build multi-user trading bots with Claude Desktop
 - ✅ Simulate different user behaviors without server restart
@@ -88,6 +119,7 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 - ✅ Operate multiple wallets from single MCP server instance
 
 ### Documentation
+
 - Added comprehensive multi-wallet section to MCP README
 - Included security considerations for privateKey usage
 - Added multi-wallet trading scenario examples
@@ -95,13 +127,16 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 ## [1.1.11] - 2025-10-01
 
 ### Changed
-- **Disabled auto-validation on startup** - Removed unnecessary GALA balance check when MCP server starts
+
+- **Disabled auto-validation on startup** - Removed unnecessary GALA balance check when MCP server
+  starts
   - Changed `autoValidate: false` to prevent startup warnings for wallets with zero balance
   - SDK still functions perfectly for all read operations (pools, prices, comments, etc.)
   - Validation only needed when performing transactions (which require GALA anyway)
   - Eliminates confusing "SDK not ready" warnings on server initialization
 
 ### Fixed
+
 - Misleading "Failed to fetch GALA balance: No balance returned" warning on startup
   - Now only validates balance when actually needed for transactions
   - Cleaner startup experience for read-only MCP usage
@@ -109,12 +144,14 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 ## [1.1.10] - 2025-10-01
 
 ### Changed
+
 - Updated SDK dependency from ^3.3.3 to ^3.3.4
 - Benefits from SDK v3.3.4 launchpad token precision fixes (18 decimals preserved)
 
 ## [1.1.9] - 2025-09-30
 
 ### Added
+
 - **Token Class Key Resolution Tool** - GalaChain utility for direct blockchain operations
   - `gala_launchpad_resolve_token_class_key` - Convert token name to GalaChain TokenClassKey
     - Returns complete TokenClassKey structure: `{collection, category, type, additionalKey}`
@@ -123,17 +160,20 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
   - AI agents can now access both utility methods for complete GalaChain integration
 
 ### Changed
+
 - Updated SDK dependency from ^3.3.1 to ^3.3.2
 - Tool count increased from 31 to 32 tools
 - Pool tools increased from 11 to 12 tools
 
 ### Documentation
+
 - SDK v3.3.2 now exports TokenClassKey types for downstream consumers
 - Both `resolveTokenClassKey()` and `resolveVaultAddress()` are official public SDK APIs
 
 ## [1.1.8] - 2025-09-30
 
 ### Added
+
 - **Spot Price Tools** - Complete price discovery for tokens (3 new tools)
   - `gala_launchpad_fetch_token_spot_price` - General DEX tokens (single or array)
     - Fetch USD spot prices for tokens like GALA, SILK, MUSIC, etc.
@@ -148,11 +188,13 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
   - AI agents can now make price-informed trading decisions
 
 ### Changed
+
 - Updated SDK dependency from ^3.3.0 to ^3.3.1
 - Tool count increased from 28 to 31 tools
 - Pool tools increased from 8 to 11 tools
 
 ### Fixed
+
 - **fetchPoolDetails response** - Now returns clean data structure
   - SDK v3.3.1 fix removes Status wrapper from response
   - MCP now receives `PoolDetailsData` directly instead of `{ Status: 1, Data: {...} }`
@@ -161,6 +203,7 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 ## [1.1.7] - 2025-09-30
 
 ### Fixed
+
 - **BREAKING:** Trading type parameter values corrected from 'token' to 'exact'
   - SDK expects `type: 'native' | 'exact'` but MCP was using `'native' | 'token'`
   - Affected tools: calculateBuyAmount, calculateSellAmount, buy, sell, calculateInitialBuy
@@ -173,11 +216,13 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
   - Affected tool: fetchPools
 
 ### Changed
+
 - Updated SDK dependency from ^3.2.0 to ^3.3.0
 - All type enums now use `Object.values(TRADING_TYPES)` and `Object.values(POOL_TYPES)`
 - Ensures MCP tool schemas always match SDK type definitions
 
 ### Benefits
+
 - Type safety via SDK constants import
 - Prevents future type mismatches
 - Auto-completion shows correct values
@@ -186,17 +231,20 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
 ## [1.1.6] - 2025-09-30
 
 ### Fixed
+
 - **postComment tool response** - Fixed MCP response validation error
   - SDK's `postComment()` returns `Promise<void>` (no data)
   - Handler was passing `undefined` to `formatSuccess()`
   - This created invalid MCP response: `{ content: [{ type: 'text', text: undefined }] }`
   - MCP SDK requires `text` field to be a string, not undefined
-  - Now returns explicit success message: `{ success: true, message: 'Comment posted successfully' }`
+  - Now returns explicit success message:
+    `{ success: true, message: 'Comment posted successfully' }`
   - Resolves: "Required" validation error on `content[0].text` path
 
 ## [1.1.5] - 2025-09-30
 
 ### Fixed
+
 - **postComment tool** - Renamed parameter from "content" to "message"
   - **BREAKING CHANGE:** Parameter renamed to avoid MCP reserved word collision
   - MCP SDK treats "content" as special parameter requiring content blocks format
@@ -205,22 +253,25 @@ This update addresses a recurring issue where AI agents forgot to pass the `reve
   - Internal: MCP parameter "message" maps to SDK parameter "content"
 
 ### Migration
+
 ```typescript
 // Before (v1.1.4 and earlier)
-gala_launchpad_post_comment({ tokenName: "anime", content: "Great token!" })
+gala_launchpad_post_comment({ tokenName: 'anime', content: 'Great token!' });
 
 // After (v1.1.5+)
-gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
+gala_launchpad_post_comment({ tokenName: 'anime', message: 'Great token!' });
 ```
 
 ## [1.1.4] - 2025-09-30
 
 ### Changed
+
 - Version bump for update verification (no code changes)
 
 ## [1.1.3] - 2025-09-30
 
 ### Fixed
+
 - **postComment tool** - Simplified to use standard MCP tool parameter pattern
   - Tool parameters should be simple JSON Schema types (string, not content blocks)
   - Content blocks format is for tool results/prompts/resources, not input parameters
@@ -230,11 +281,13 @@ gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
 ## [1.1.2] - 2025-09-30 [DEPRECATED]
 
 ### Fixed (Incorrect Approach)
+
 - ~~postComment tool schema - Used overly complex oneOf schema~~ (reverted in 1.1.3)
 
 ## [1.1.1] - 2025-09-30
 
 ### Fixed
+
 - **postComment tool** - Fixed content format handling for MCP protocol
   - Now correctly extracts text from MCP content blocks array
   - Handles both direct string and MCP content blocks format
@@ -244,28 +297,38 @@ gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
   - Ensures all 28 MCP tools are functional
 
 ### Dependencies
+
 - Updated `@gala-chain/launchpad-sdk` from ^3.1.0 to ^3.2.0
 
 ## [1.1.0] - 2025-09-30
 
 ### Added
-- **AI-AGENT-PATTERNS.md** - Comprehensive AI agent guide with common workflows, gotchas, and best practices
+
+- **AI-AGENT-PATTERNS.md** - Comprehensive AI agent guide with common workflows, gotchas, and best
+  practices
 - **CONSTRAINTS-REFERENCE.md** - Technical reference for SDK-to-MCP constraint mappings
-- **Central constraints utility** (`src/utils/constraints.ts`) - Single source of truth for pagination limits
+- **Central constraints utility** (`src/utils/constraints.ts`) - Single source of truth for
+  pagination limits
 - **AI-friendly error messages** - Enhanced error handler with contextual tips for common errors
 - **Workflow documentation** - Added detailed workflow comments to trading and creation tools
 
 ### Fixed
+
 - **BREAKING:** Fixed constraint mismatches between MCP tool schemas and SDK validation
   - `fetchTrades`: Reduced limit from 100 → 20 (matches `TRADE_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
   - `fetchTokensHeld`: Reduced limit from 100 → 20 (matches `USER_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
-  - `fetchTokensCreated`: Reduced limit from 100 → 20 (matches `USER_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
-  - `fetchComments`: Reduced limit from 100 → 50 (matches `COMMENT_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
+  - `fetchTokensCreated`: Reduced limit from 100 → 20 (matches
+    `USER_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
+  - `fetchComments`: Reduced limit from 100 → 50 (matches
+    `COMMENT_CONSTRAINTS.PAGINATION.MAX_LIMIT`)
   - `fetchPools`: Verified limit 100 is correct (matches `PAGINATION_CONSTRAINTS.MAX_LIMIT`)
 
 ### Improved
-- **expectedAmount parameter documentation** - Added critical warnings and workflow examples to prevent confusion
-  - Clarified that `expectedAmount` must be `result.amount` from calculate functions, NOT input amount
+
+- **expectedAmount parameter documentation** - Added critical warnings and workflow examples to
+  prevent confusion
+  - Clarified that `expectedAmount` must be `result.amount` from calculate functions, NOT input
+    amount
   - Added step-by-step workflow examples with actual values
   - Included common mistake examples with corrections
 - **Type parameter clarity** - Enhanced 'native' vs 'token' documentation with use-case examples
@@ -279,12 +342,14 @@ gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
   - Type parameter confusion (with usage examples)
 
 ### Documentation
+
 - Enhanced inline comments throughout tool schemas linking to SDK constraint sources
 - Added comprehensive API documentation for AI agents
 - Created quick reference cards for common patterns
 - Documented all constraint mappings with source file line numbers
 
 ### Developer Experience
+
 - AI agents can now self-diagnose and fix common errors using enhanced error messages
 - Reduced trial-and-error cycles with upfront workflow documentation
 - Improved discoverability with centralized documentation
@@ -293,6 +358,7 @@ gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
 ## [1.0.4] - 2025-09-18
 
 ### Added
+
 - Initial release with 28 MCP tools for Gala Launchpad
 - Support for trading, balance queries, pool management, and token creation
 - Integration with `@gala-chain/launchpad-sdk` v3.1.0
@@ -302,17 +368,21 @@ gala_launchpad_post_comment({ tokenName: "anime", message: "Great token!" })
 **Migration Guide v1.0.4 → v1.1.0:**
 
 ⚠️ **Breaking Changes:**
-- If your code uses `limit` values > 20 for `fetchTrades`, `fetchTokensHeld`, or `fetchTokensCreated`, reduce to ≤ 20
+
+- If your code uses `limit` values > 20 for `fetchTrades`, `fetchTokensHeld`, or
+  `fetchTokensCreated`, reduce to ≤ 20
 - If your code uses `limit` values > 50 for `fetchComments`, reduce to ≤ 50
 - These limits now match the SDK's actual validation constraints
 
 ✅ **Recommended Actions:**
+
 1. Update any hardcoded `limit` values to respect new constraints
 2. Review trading workflows to ensure `expectedAmount` uses calculated `result.amount`
 3. Consult `docs/AI-AGENT-PATTERNS.md` for updated best practices
 4. Leverage enhanced error messages for faster debugging
 
 📚 **New Documentation:**
+
 - `docs/AI-AGENT-PATTERNS.md` - AI agent guide
 - `docs/CONSTRAINTS-REFERENCE.md` - Constraint mappings
 - `src/utils/constraints.ts` - Constraint constants

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gala-chain/launchpad-mcp-server",
-  "version": "1.2.13",
+  "version": "1.2.15",
   "description": "MCP server for Gala Launchpad SDK with 37 tools - AI agents can interact with Gala Launchpad",
   "main": "dist/index.js",
   "bin": {
@@ -56,8 +56,14 @@
     "registry": "https://registry.npmjs.org/"
   },
   "dependencies": {
-    "@gala-chain/launchpad-sdk": "^3.6.7",
+    "@gala-chain/launchpad-sdk": "^3.7.1",
     "@modelcontextprotocol/sdk": "^0.5.0",
+    "ethers": "^6.15.0",
+    "@gala-chain/api": "^2.4.3",
+    "@gala-chain/connect": "^2.4.3",
+    "socket.io-client": "^4.8.1",
+    "axios": "^1.12.2",
+    "bignumber.js": "^9.1.2",
     "zod": "^3.25.76",
     "zod-to-json-schema": "^3.24.6"
   },