@livx.cc/agentx 0.99.84 → 0.99.86

package/README.md

@@ -64,7 +64,7 @@ await fs.createDir('/src');
 await fs.writeFile('/src/x.ts', 'export const add = (a,b) => a - b;\n');
 
 const ai = new AIClient({ apiKeys: { anthropic: process.env.ANTHROPIC_API_KEY } });
-const res = await new Agent({ ai, fs, model: 'anthropic/claude-sonnet-4-6' })
+const res = await new Agent({ ai, fs, model: 'anthropic/sonnet' })
   .run('Fix the add bug in /src/x.ts');
 
 console.log(res.finishReason, await fs.readFile('/src/x.ts'));
@@ -77,7 +77,7 @@ console.log(res.finishReason, await fs.readFile('/src/x.ts'));
 - **`Edit`** — exact unique-substring replace, with a read-before-edit staleness guard.
 - **`Grep`/`Glob`/`Write`/`MultiEdit`** — structured, typed results straight from the VFS (no `bash` parsing). The selectable tool set the self-evolution loop mutates over.
 - **`TodoWrite`** — a planning scratchpad; **`Task`** — spawn a depth-limited child agent over the VFS (`subagents: true`); **`SlashCommand`** — reusable prompt templates from `<dir>/*.md` (`commandsDir`); plus a real **MCP client** (`src/mcp.client.ts`, node-only — stdio/HTTP JSON-RPC handshake + discovery) that feeds the edge-safe **MCP adapter** (`mcpToolsToAgentTools`), so any MCP server's tools become agent tools.
-- **`WebFetch`/`WebSearch`** — fetch a URL as readable text, or search the web. **Keyless by default** (WebSearch uses DuckDuckGo; `provider: 'auto'` upgrades by key presence — Firecrawl `FIRECRAWL_API_KEY` > Tavily `TAVILY_API_KEY`) and **auto-enabled in the CLI**. Factory-built with an injectable `fetch`, so they stay edge-portable and testable. (In the library they're opt-in by name: `tools: [...,'WebSearch']`.) **`WebSearchAnthropic`** is a separate provider-pinned tool — Anthropic's native search relayed through a cheap model (`claude-haiku-4-5`); CLI-default-on when an Anthropic key resolves, API-billed per call, opt out via `anthropicWebSearch: false`.
+- **`WebFetch`/`WebSearch`** — fetch a URL as readable text, or search the web. **Keyless by default** (WebSearch uses DuckDuckGo; `provider: 'auto'` upgrades by key presence — Firecrawl `FIRECRAWL_API_KEY` > Tavily `TAVILY_API_KEY`) and **auto-enabled in the CLI**. Factory-built with an injectable `fetch`, so they stay edge-portable and testable. (In the library they're opt-in by name: `tools: [...,'WebSearch']`.) **`WebSearchAnthropic`** is a separate provider-pinned tool — Anthropic's native search relayed through a cheap model (the newest Haiku); CLI-default-on when an Anthropic key resolves, API-billed per call, opt out via `anthropicWebSearch: false`.
 - **Oversized-output pagination** — any tool result over a byte ceiling (`maxToolResultBytes`, default 60k) is cropped to page 1 with a marker (refine the query / read further), so one big `Grep`/`Read`/MCP/web result can't blow the context window. In the CLI (**on by default**; `--no-scratch` to disable) the full output instead spills **losslessly** to a **scratch** file and the model recovers specifics via `Grep`/`Read` or **`Ask`** — a cheap, context-isolated peek that returns just the answer (the raw blob never re-enters context).
 
 ## Agentic subsystems

package/dist/cli.d.ts

@@ -133,6 +133,7 @@ interface Args {
     help: boolean;
     version: boolean;
     duplex: boolean;
+    duplexV2?: boolean;
     voiceModel?: string;
     thinkModel?: string | false;
     voice: boolean;