@mu-agent/mu 0.2.4 → 0.3.1

package/README.md

@@ -56,12 +56,13 @@ The CLI loads the coding profile by default. It includes:
 - read-before-write checks and diff previews for approvals;
 - shadow-git workspace checkpoints with `/undo`, `/redo`, `/fork`, and `/diff`;
 - durable sessions, automatic context compaction, and Markdown transcript export;
-- managed `task` delegation and coding-specific `search`, `counsel`, and `recall` subagents.
+- managed `task` delegation when you ask for subagents, and coding-specific `counsel` and
+  `recall` subagents.
 
 Explicitly ask **“Ask Recall why we rejected SQLite”** to investigate retained project
 sessions, including compacted history and alternate branches. Recall has only read-only
-history tools, no workspace or network tools, and cites session/entry references. It uses
-Search's same-provider model routing with the current model as fallback. Explicit-only
+history tools, no workspace or network tools, and cites session/entry references. It runs on
+the current model at `low` reasoning unless pinned under `/subagents`. Explicit-only
 invocation is a parent-model instruction, not a classifier or runtime authorization gate.
 Separate project directories, including other clones/worktrees, are not automatically joined.
 
@@ -79,15 +80,17 @@ user's repository.
 |---|---|
 | `/login`, `/logout` | Add or remove credentials stored by Mu |
 | `/model` | Select the active model |
+| `/thinking [level]` | Set the active model's reasoning level (same as ctrl+t); bare, it lists the levels |
 | `/permissions` | Change the current permission mode |
 | `/resume`, `/new`, `/rename` | Open, create, or name conversations |
 | `/compact [focus]` | Compact the active model context |
 | `/undo`, `/redo`, `/fork`, `/diff` | Navigate the session and workspace history |
 | `/export [path.md]` | Export the complete active branch |
 | `/btw [question]` | Open an ephemeral read-only side conversation |
+| `/skills` | List skills and set each on, manual or off |
 | `/instructions [reload]`, `/reload` | Inspect or reload repository instructions |
-| `/subagents` | List subagents, their roles, and how to invoke them |
-| `/cost` | Show session token and cost totals |
+| `/subagents` | List subagents, the model and reasoning level each runs on, and how to invoke them; choose a different model or level for any of them (coding profile) |
+| `/cost` | Show spend by source and model, and what fills the context |
 | `/keybindings` | Show Mu-specific keyboard controls |
 
 During a run, Enter steers the current turn and Tab queues a follow-up. `Ctrl+O` opens
@@ -98,15 +101,19 @@ are still recorded as session context.
 ## Providers
 
 Account sign-in is available for OpenAI Codex/ChatGPT, GitHub Copilot, Kimi Code,
-OpenRouter, and xAI. API-key routes include Anthropic, OpenAI, Google, Z.AI Coding Plan,
-Qwen Token Plan, and providers supported by Mu's shared OpenAI-compatible and
-Gemini-compatible transports. Anthropic subscription OAuth is not supported.
+OpenRouter, and xAI. API-key routes include Anthropic, OpenAI, Google, Inception Labs,
+Z.AI Coding Plan, Qwen Token Plan, and providers supported by Mu's shared OpenAI-compatible
+and Gemini-compatible transports. Anthropic subscription OAuth is not supported.
 
 OpenAI API-key models use the `openai/*` namespace. ChatGPT-plan models use
 `openai-codex/*`; the credentials are separate and may coexist. Both routes receive
 provider-hosted web search automatically. Other providers currently run without web
 search.
 
+For Inception Labs, use `/login` → API key → Inception Labs or set `INCEPTION_API_KEY`,
+then run `mu --model inception/mercury-3`. Mercury 2.5 and Mercury 2 are also supported;
+`/thinking` selects `instant`, `low`, `medium`, or `high` reasoning.
+
 A local `llama-server` is discovered at `http://127.0.0.1:8000` and exposed as
 `llama-cpp/<alias>`. Set `LLAMA_CPP_BASE_URL` or `LLAMA_CPP_API_KEY` when needed.
 
@@ -137,8 +144,10 @@ The CLI also discovers:
 
 - TypeScript extensions in `~/.mu/extensions` and `.mu/extensions`;
 - Markdown slash commands in `~/.mu/commands` and `.mu/commands`;
-- skills in `~/.mu/skills` and `.mu/skills`;
-- MCP servers from user and project `.mu/config.json`.
+- skills in `.mu/skills`, `.agents/skills`, `.claude/skills`, `.codex/skills` and
+  `.opencode/skills` (project and home);
+- MCP servers from user and project `.mu/config.json`, and those configured for Claude Code,
+  Codex, Cursor, opencode, Gemini and VS Code.
 
 ```json
 {
@@ -146,14 +155,15 @@ The CLI also discovers:
     "docs": {
       "command": "bunx",
       "args": ["-y", "your-mcp-server"],
-      "env": { "SERVER_TOKEN": "..." }
+      "env": { "SERVER_TOKEN": "${DOCS_TOKEN}" }
     }
   }
 }
 ```
 
-MCP tools are named `mcp_<server>_<tool>` and use the same permission, event, and abort
-paths as local tools.
+The model opens a server to see its tools and calls them through `mcp_call`, which uses the
+same permission, event, and abort paths as local tools. Project-defined servers need
+`/mcp approve <name>`; `/mcp` and `/skills` show what was found.
 
 ## TypeScript SDK
 
@@ -185,7 +195,7 @@ console.log(result.reason, result.usage);
 
 `createAgent()` adds managed `task` delegation. With `profile: "coding"`, it also loads
 the shipped coding tools, project instructions, permissions, checkpoints, background
-runtime, and `search`/`counsel`/`recall` specialists. Recall can access only history in the
+runtime, and `counsel`/`recall` specialists. Recall can access only history in the
 configured session store; SDK agents use an in-memory store by default.
 
 ```ts