dsh-mcp-lens
Cut MCP context and API cost in DeepSeek Harness with two model-facing tools, lazy discovery, and exact-schema calls
AI 분석
核心用途是优化 MCP（Model Context Protocol）的上下文占用与 API 开销。适合频繁使用 MCP 工具、希望降低 Token 消耗和费用的用户。
설치
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:labmimors/dsh-mcp-lensConfiguration reference
Most users only need servers, cachePath, allowTools, and denyTools. The remaining fields already have bounded defaults:
Show all bounded defaults
| Field | Default | Purpose |
|---|---|---|
catalogTtlMs | 86400000 | Refresh a catalog after 24 hours |
idleDisconnectMs | 300000 | Close an idle server after 5 minutes |
connectTimeoutMs | 30000 | Connection deadline |
callTimeoutMs | 60000 | Tool-call deadline |
discoveryTimeoutMs | 30000 | Whole paginated discovery deadline |
maxDiscoveryPages | 1000 | Maximum pages per discovery |
maxToolsPerServer | 10000 | Maximum tools accepted from one server |
maxBytesPerTool | 1048576 | Maximum projected metadata bytes per tool |
maxTotalCatalogBytes | 67108864 | Maximum total catalog/cache bytes |
maxHttpResponseBytes | 16777216 | Maximum streamed HTTP response bytes |
maxCursorBytes | 4096 | Maximum UTF-8 pagination cursor bytes |
searchLimitDefault | 5 | Default search results |
searchLimitMax | 10 | Maximum search results |
See the shipped cordis.patch.yml for the canonical defaults.