command-code 0.52.5 → 1.0.0

package/dist/bundled/command-code-knowledge/reference/plan-mode.md

@@ -0,0 +1,101 @@
+<!-- GENERATED FILE — do not edit. Source: packages/docs/src/app/core-concepts/plan-mode/page.mdx. Regenerate: pnpm generate:knowledge -->
+
+# Plan Mode and Auto-Accept Mode
+
+Command Code separates reasoning from execution using permission modes:
+
+- **Plan Mode**: Explore and plan without executing file changes
+- **Auto-Accept Mode**: Implement changes without confirmation prompts
+
+`Shift + Tab` is your CLI shortcut to switch between plan and auto-accept modes.
+
+---
+
+## Plan Mode
+
+Plan mode helps you with structured thinking. It can:
+
+- Read your codebase
+- Search files
+- Analyze architecture
+- Discuss tradeoffs
+- Propose implementation plans
+
+It **cannot**:
+
+- Modify files
+- Run shell commands
+- Apply patches
+
+We recommend Plan mode to:
+
+- Explore unfamiliar codebases
+- Debug complex issues
+- Design multi-file features
+- Evaluate architecture decisions
+- Identify edge cases early
+- Review sensitive or critical flows
+
+---
+
+## Auto-Accept Mode
+
+Auto-Accept mode is for fast implementation and execution.
+
+- Modifies files directly
+- Applies patches without confirmation
+- Executes shell commands
+- Proceeds step-by-step automatically
+
+We recommend Auto-Accept mode when:
+
+- The approach is already clear
+- You are iterating quickly
+- You trust the implementation path
+
+Command Code creates **Checkpoints** before modifying files, so you can rewind at any time.
+
+Learn how to use [checkpoints](./checkpoints.md) in Command Code.
+
+---
+
+## Default Mode
+
+Default mode sits between the Plan and Auto-Accept modes.
+
+- File edits require confirmation
+- Commands require approval
+
+Use the Default mode when you want controlled iteration without fully separating planning and execution.
+
+---
+
+## Suggested workflow with modes
+
+1. Start in **Plan Mode**
+2. Explore relevant files
+3. Discuss approach and edge cases
+4. Switch to **Auto-Accept Mode** if you're ready to implement
+5. Rewind with Checkpoints if needed
+
+---
+
+## Example use cases for choosing the right mode
+
+|            Use Case            | Recommended Mode |
+| ------------------------------ | ---------------- |
+| New feature with unclear scope | Plan             |
+| Debugging complex issue        | Plan             |
+| Multi-file refactor            | Plan             |
+| Security review                | Plan             |
+| Small bug fix                  | Auto-Accept      |
+| Routine refactor               | Auto-Accept      |
+| Running tests and adjusting    | Auto-Accept      |
+| Quick typo fix                 | Auto-Accept      |
+
+---
+
+## Next Steps
+
+- Use **Plan Mode** to outline a change, then switch to **Auto-Accept** to implement it
+- Join our [Discord community](https://commandcode.ai/discord) for support

package/dist/bundled/command-code-knowledge/reference/product-help.md

@@ -0,0 +1,421 @@
+<!-- GENERATED FILE — do not edit. Source: packages/harness/src/tools/get-command-code-knowledge/knowledge-content.ts. Regenerate: pnpm generate:knowledge -->
+
+# Command Code — Product Knowledge
+
+## KEYBOARD SHORTCUTS
+- Shift+Tab: Toggle mode (default → auto-accept → plan)
+- Ctrl+O: Toggle expanded tool output
+- Alt+P: Quick model switch (Option+P on macOS)
+- Ctrl+G: Open input in external editor ($EDITOR)
+- Press Esc twice: Rewind to previous checkpoint
+- /: Open command menu
+
+## SLASH COMMANDS (type these in the chat input)
+- /init: Initialize AGENTS.md for this project
+- /import [claude|codex|cursor|pi|opencode|gemini]: Import your setup (skills, agents, commands, MCP, memory) from another coding agent
+- /goal [<objective>|clear|status]: Set an objective for the agent to work towards
+- /memory: Manage Command Code memory
+- /resume: Resume a past conversation
+- /sessions: Resume a past conversation (alias of /resume)
+- /fork [name]: Fork the conversation into a new session
+- /worktree [name|list|remove <name>]: Create, list, or switch isolated git worktrees
+- /clone: Clone the current branch into a new session and switch to it
+- /rename [name]: Rename the current session
+- /name [name]: Alias of /rename; bare /name shows the current name
+- /rewind: Restore to a previous checkpoint (Press Esc twice)
+- /tree: Browse the session tree and jump to any point in it
+- /clear (new): Start a new session with empty context; previous stays on disk, resumable with /resume (/new is the same command)
+- /theme: Switch between dark and light themes
+- /share [gist]: Share conversation — /share [gist [html|jsonl|md]]
+- /unshare: Stop sharing conversation
+- /taste: Manage Taste learning and usage
+- /learn-taste: Learn taste from sessions with other coding agents (Claude Code, Cursor, etc)
+- /skills: Browse and open agent skills
+- /agents: Manage agent configurations
+- /design: Design partner: audit, build, compose, and ship UI — type /design to browse modes
+- /mcp: Manage MCP server connections
+- /model: Switch between Command Code models
+- /effort: Set reasoning effort for the current model
+- /provider: Select AI provider (Command Code or Anthropic)
+- /compact: Compact the conversation history
+- /compact-mode: Select a compact mode to compact sessions
+- /config: Search and change settings, including which model runs each built-in task
+- /context: Show context window usage and breakdown
+- /ide: Connect IDE to share your open file and selected lines
+- /terminal-setup: Setup VSCode keybindings
+- /login: Log in to Command Code
+- /logout: Log out of Command Code
+- /courses: Open Command Code courses in your browser
+- /feedback [title]: Share feedback or report bugs (optional title)
+- /trace: Copy the current trace id; required for support debugging
+- /session-file: Show the current session id and path to the on-disk session file
+- /copy: Copy the last response to the clipboard
+- /export [path]: Export session — /export [html|jsonl|md] or a file path
+- /session: Show session info: id, model, messages, context
+- /changelog: See what's new in Command Code
+- /hotkeys: Show all keyboard shortcuts
+- /todos: Manage the session todo list — x to remove an item, c to clear
+- /plan [task]: Enter plan mode; `/plan <task>` plans that task
+- /mode [name]: Show or switch the permission mode — /mode [default|auto-accept|plan]
+- /mode:default: Switch to default mode (prompt before edits/commands)
+- /mode:auto-accept: Switch to auto-accept mode (accept edits automatically)
+- /mode:plan: Switch to plan mode (read-only, no side effects)
+- /plans [name]: Browse, review, and annotate saved plans
+- /plan-review: Review this session's latest plan
+- /review [pr]: Review a pull request (optional PR number)
+- /pr-comments: Fetch all PR comments for current branch
+- /add-dir: Manage additional directory scope
+- /status: Show comprehensive environment status
+- /usage: Display credits, plan, and usage metrics
+- /upgrade: Open billing page to upgrade your plan
+- /extra: Open billing page to buy on-demand credits
+- /update: Update Command Code to the latest version
+- /reload: Restart Command Code and resume this session (applies a staged update)
+- /help: Display help information
+- /exit: Exit Command Code
+- /quit: Exit Command Code (alias of /exit)
+
+## CLI SUBCOMMANDS
+- cmd info: Display system information
+- cmd status: Show authentication status
+- cmd help: Display help information
+- cmd whoami: Show current user
+- cmd update: Update Command Code to the latest version
+- cmd feedback [title]: Share feedback or report bugs (optional title)
+- cmd taste: Manage taste learning packages
+- cmd taste learn <source>: Learn taste from a local repository or GitHub repo
+- cmd learn-taste: Learn command structure from repositories
+- cmd mcp: Manage MCP (Model Context Protocol) servers
+- cmd skills: Manage skills from GitHub repositories
+- cmd mods: Manage mods (loadable extensions) from npm, git, or local paths
+- cmd login: Login with Command Code account
+- cmd logout: Log out of Command Code
+
+## CLI EXAMPLES
+- cmd: Start interactive session
+- cmd "fix the login bug": Start with a task
+- cmd -c: Continue last conversation
+- cmd -r: Resume a past session
+- cmd --resume "auth refactor": Resume a named session
+- cmd -p "your query": Run non-interactive query
+- cmd taste learn .: Learn taste from the current repository
+- cmd taste learn owner/repo: Learn taste from a GitHub repository
+- cmd --add-dir ../shared: Start with additional directory scope
+- cmd info: Show system information
+- cmd feedback "title": Open feedback form with title
+
+## CLI OPTIONS (flags when launching cmd)
+- cmd: Start interactive session
+- cmd "message": Start with initial message
+- -r, --resume [name]: Resume a conversation by id or name (use quotes for multi-word names), or pick from history
+- -c, --continue: Continue the last conversation
+- --fork-session: With --resume/--continue, fork the session into a new one (original left untouched)
+- --session <path|id>: Resume a session by transcript path (.jsonl) or a unique session-id prefix
+- --no-session: Don't persist this session to disk (in-memory only)
+- -n, --name <name>: Set the session display name
+- -t, --trust: Auto-trust project (skip initial permission prompt)
+- -p, --print [query]: Run in non-interactive mode, output response and exit
+- --max-turns <number>: Cap conversation turns in -p mode (default 100; exit 8 on cap-hit)
+- --output-format <format>: -p output: text (default) or json (NDJSON event stream + final result line)
+- -m, --model <model>: Run on a specific model this session
+- --effort <level>: Set reasoning effort for the session (e.g. low, medium, high) — depends on the model
+- --theme <theme>: Set the color theme (dark or light)
+- --config <key=value>: Set any setting headlessly, e.g. --config theme=dark (repeatable) — the CLI form of /config
+- --list-models: List the models available for use
+- --effort <level>: Set reasoning effort (low, medium, high — per model)
+- --theme <theme>: Set the color theme (dark or light)
+- --config <key=value>: Set a setting headlessly, e.g. --config theme=dark (repeatable)
+- --plan: Start in plan mode
+- --permission-mode <mode>: Set permission mode (standard, plan, auto-accept)
+- --auto-accept: Start in auto-accept mode
+- --yolo: Bypass all permission prompts (alias for --dangerously-skip-permissions)
+- --add-dir <directory>: Add directory to workspace context
+- -w, --worktree [name]: Run in an isolated managed worktree (name, path, or #PR; generated when omitted)
+- --mod <path>: Load a mod file or directory for this session (repeatable)
+- --mod-option <name=value>: Set a mod-declared flag value (repeatable)
+- --skill <path>: Load extra skills from a path (a skill directory or a directory of skills); repeatable
+- --no-skills: Skip skill discovery; paths given via --skill still load
+- --skip-onboarding: Skip taste onboarding (for automated runs)
+- --ide-setup: Connect IDE to share your open file and selected lines
+- --no-auto-update: Disable automatic background updates for this run
+- -v, --version: Output the version number
+- -h, --help: Display this help message
+
+## TASTE SYSTEM
+Taste is powered by the meta neuro-symbolic AI model taste-1 with continuous reinforcement learning (RL). It combines reasoning with neural intuition to learn your coding preferences.
+
+How taste works:
+- Learns from you — every accept, reject, and edit becomes a signal
+- Thinks like you — learns patterns and micro-decisions you'd never document
+- Grows with you — continuous learning loop that never goes stale
+
+Enable taste: Use /taste in a session or run npx taste from the command line.
+
+Taste packages (three types):
+- Project: Stored in .commandcode/taste/ — learnings unique to this codebase
+- Global: Stored in ~/.commandcode/taste/ — personal taste across all projects (use -g flag)
+- Remote: Stored at commandcode.ai/username/taste — team sharing, backup, sync across machines
+
+File structure:
+  .commandcode/taste/
+  ├── taste.md (main taste file)
+  ├── cli/taste.md
+  ├── typescript/taste.md
+  └── architecture/taste.md
+
+Key commands:
+- npx taste push --all — push entire project taste to remote
+- npx taste pull username/project-name — pull taste from remote
+- npx taste push cli -g — push package to global
+- npx taste pull cli -g — pull package from global
+- npx taste list — list available packages
+- npx taste lint — validate package format
+- npx taste open — open packages in editor
+
+Privacy: Taste processing runs on your codebase and stores learning data locally only.
+
+## KEY FEATURES
+- Taste System: Automatically learns coding preferences from interactions and applies them consistently
+- Skills: User-defined knowledge modules for specialized tasks (stored in .commandcode/skills/ or .agents/skills/)
+- Agents: Custom agent configurations for different workflows (stored in .commandcode/agents/)
+- Memory: Project-specific instructions via COMMANDCODE.md files
+- MCP: Model Context Protocol server connections for extended tool capabilities
+- Plan Mode: Structured exploration and planning before implementation [Shift+Tab to toggle]
+- Conversation Management: Resume, continue, share, and rewind conversations
+- PR Review: Built-in pull request review capabilities (/review command)
+- Taste Learning: Learnings appear inline in the conversation as TASTE blocks while you work
+- Model Switching: Switch between models with Alt+P (Option+P on macOS)
+
+## EXTENDING COMMAND CODE
+Command Code is extended with skills, agents, MCP servers, and project memory — all discovered from files in your project or home directory, no restart required.
+
+SKILLS (reusable playbooks): a directory containing a SKILL.md file, discovered from:
+- .commandcode/skills/<name>/SKILL.md (project)
+- ~/.commandcode/skills/<name>/SKILL.md (personal, across all projects)
+- .agents/skills/<name>/SKILL.md (compat location, project or personal)
+
+Minimal SKILL.md:
+---
+name: commit-messages
+description: Write a conventional commit message from staged changes. Use when the user asks to commit or write a commit message.
+---
+
+# Commit Messages
+1. Run `git diff --staged` to see what changed.
+2. Write a one-line conventional-commit summary (type(scope): subject).
+3. Add a short body only if the change needs explaining.
+
+Frontmatter fields: name (required — lowercase letters/numbers/hyphens only, must match the directory name), description (required, ≤1024 chars — this is what the model matches against), allowed-tools (optional — a space-delimited allowlist, e.g. "Bash(git:*) Read"), argument-hint (optional — shown when the skill is invoked with arguments), disable-model-invocation (optional true/false — true stops the model from auto-invoking it; /name still works). Invoke with /<name>, or let the model invoke it automatically when the description matches the task. Manage installed skills with /skills.
+
+AGENTS (sub-agents for the task tool): a single Markdown file at .commandcode/agents/<name>.md (project) or ~/.commandcode/agents/<name>.md (personal). Full authoring guide — file format, model pinning, and the create-an-agent-on-request recipe — is in the CUSTOM AGENTS section (topic: agents).
+
+MCP SERVERS (external tools and data): add one with the CLI —
+  cmd mcp add --transport http notion https://mcp.notion.com/mcp
+  cmd mcp add github -- npx -y @modelcontextprotocol/server-github
+or hand-write .mcp.json in the project root:
+  {
+    "mcpServers": {
+      "github": {
+        "command": "npx",
+        "args": ["-y", "@modelcontextprotocol/server-github"]
+      }
+    }
+  }
+Scopes, lowest to highest precedence: user (~/.commandcode/mcp.json) < project (<project>/.mcp.json, checked in and shared) < local (~/.commandcode/projects/<slug>/mcp.json, private to this machine, not checked in). Manage connections with /mcp.
+
+PROJECT MEMORY (AGENTS.md): the project-root AGENTS.md holds conventions the agent should always follow (build/test commands, code style, architecture notes). A nested <dir>/AGENTS.md layers in directory-specific rules for files under that directory (nearest wins). An @path reference inlines another file's content (skipped inside fenced code blocks). Run /init to scaffold one, /memory to edit it.
+
+## CUSTOM AGENTS (sub-agents)
+Custom agents are specialized sub-agents the main session delegates to via the task tool. Each is ONE Markdown file — no build step, no restart; the registry re-reads disk every turn, so a new agent is usable on the next message.
+
+Locations (project wins a name clash):
+- .commandcode/agents/<name>.md — this project
+- ~/.commandcode/agents/<name>.md — personal, all projects
+
+Complete agent file:
+---
+name: "changelog-writer"
+description: "Writes and updates CHANGELOG.md entries from recent commits. Use after merging changes or when the user asks for a changelog."
+tools: "read_file, grep, glob, shell_command, edit_file"
+model: "moonshotai/Kimi-K2.5"
+---
+
+You write crisp changelog entries. Read the recent commits, group them by type (features, fixes), and write terse one-liners under the right heading in CHANGELOG.md.
+
+Frontmatter fields:
+- name (required) — lowercase-kebab-case; becomes the subagent_type the task tool dispatches on. Reserved (do not use): explore, plan, review, general.
+- description (required) — when to delegate to this agent; the orchestrator picks agents by this text.
+- tools (optional) — "*" for every tool, or a comma/space-delimited list of tool names; OMITTED MEANS NO TOOLS. Grant the minimum the role needs (read-only for reviewers/analysts; add edit_file/shell_command only for agents that must change things).
+- model (optional) — pins the model this agent ALWAYS runs on, independent of the session's /model. Omit it (or write "inherit") to follow the session model. Pinned agents keep their own prompt cache.
+
+Valid model ids (the /model catalog — use these EXACT ids):
+- claude-sonnet-5
+- claude-sonnet-4-6
+- claude-fable-5
+- claude-opus-4-8
+- claude-opus-4-7
+- claude-haiku-4-5-20251001
+- gpt-5.6-sol
+- gpt-5.6-terra
+- gpt-5.6-luna
+- gpt-5.5
+- gpt-5.4
+- gpt-5.3-codex
+- gpt-5.4-mini
+- MiniMaxAI/MiniMax-M3-Free
+- moonshotai/Kimi-K3
+- thinkingmachines/inkling
+- deepseek/deepseek-v4-pro
+- deepseek/deepseek-v4-flash
+- moonshotai/Kimi-K2.7-Code
+- moonshotai/Kimi-K2.7-Code-Highspeed
+- moonshotai/Kimi-K2.6
+- moonshotai/Kimi-K2.5
+- zai-org/GLM-5.2
+- zai-org/GLM-5.2-Fast
+- zai-org/GLM-5.1
+- zai-org/GLM-5
+- MiniMaxAI/MiniMax-M3
+- MiniMaxAI/MiniMax-M2.7
+- MiniMaxAI/MiniMax-M2.5
+- xiaomi/mimo-v2.5-pro
+- xiaomi/mimo-v2.5
+- Qwen/Qwen3.6-Max-Preview
+- Qwen/Qwen3.6-Plus
+- Qwen/Qwen3.7-Max
+- Qwen/Qwen3.7-Plus
+- stepfun/Step-3.7-Flash
+- stepfun/Step-3.5-Flash
+- tencent/Hy3
+- google/gemini-3.6-flash
+- google/gemini-3.5-flash
+- google/gemini-3.5-flash-lite
+- google/gemini-3.1-flash-lite
+- sakana/fugu-ultra
+- xai/grok-4.5
+- meta/muse-spark-1.1
+- nvidia/nemotron-3-ultra-550b-a55b
+- poolside/laguna-s-2.1-free
+A BYO-provider id (from a configured custom provider) also works and is passed through as-is.
+
+WHEN THE USER ASKS YOU TO CREATE AN AGENT (e.g. "make a changelog agent with a cheap model"):
+1. Do NOT search the codebase for how — this section is the authoritative spec.
+2. Derive a lowercase-kebab-case name (not a reserved name), a delegation-worthy description, and a short, focused system prompt for the role.
+3. Pick the minimal tool set for the role.
+4. If the user named a model (or a tier like "cheap"/"fast"/"deep reasoning"), map it to an EXACT id from the list above and set model:; otherwise omit model: so the agent inherits the session model. Never invent a model id.
+5. Write the file to .commandcode/agents/<name>.md (project) unless the user asked for a personal agent (~/.commandcode/agents/<name>.md).
+6. Confirm: the agent is live on the next message — the user can also see it under /agents, and the task tool can now dispatch to it by name.
+
+To modify an existing agent, edit its file in place (same fields). To remove one, delete the file.
+
+## MODS (loadable plugins — the ModApi)
+Mods are loadable plugins written against the ModApi — a TypeScript file that Command Code discovers on disk and loads onto its agent loop. This is the plugin layer above hooks/skills: a mod can add tools the model calls, slash commands, mutating lifecycle hooks, event observers, typed-input interception, custom feed rendering, flags, and model providers. Command Code's own built-in features (providers, session titling, the update notice) are written as mods against this same API.
+
+A mod default-exports a factory that receives the API bound as `cmd`:
+
+import type {ModApi} from '@commandcode/harness';
+export default function (cmd: ModApi) {
+  cmd.addCommand({name: 'hello', handler: () => ({message: 'hi from a mod'})});
+}
+
+Save it at ~/.commandcode/mods/<name>.ts (personal) or <project>/.commandcode/mods/<name>.ts (project, trust-gated). It loads next session, or test now with `cmd --mod ./<name>.ts`. No build step — jiti compiles at load. Manage packages with `cmd mods add|remove|list|update`.
+
+Registration verbs (each returns a Disposable): cmd.addTool, cmd.addCommand, cmd.addFlag, cmd.addProvider, cmd.addRenderer, cmd.on(event, handler), and cmd.hooks({...}). THE ONE RULE: cmd.hooks mutates (block/rewrite a tool via beforeToolCall/afterToolCall, add to the prompt via appendSystemPrompt, rewrite typed input via transformInput, post-turn work via onRunEnd); cmd.on only observes.
+
+To help a developer build one: use the bundled "mod-builder" skill (/mod-builder or let the model invoke it) — it has runnable single-file examples for each capability plus the full mods reference under its reference/ directory (generated from the public docs at https://commandcode.ai/docs/mods).
+
+## FREQUENTLY ASKED QUESTIONS
+Q: What is Command Code?
+A: Command Code is the first coding agent that automatically learns your taste of writing great code.
+
+Q: What is taste-1?
+A: taste-1 is our meta neuro-symbolic AI model with continuous reinforcement learning (RL).
+
+Q: Is Command Code free?
+A: Command Code offers both free and premium plans. Check https://commandcode.ai/pricing for details.
+
+Q: How do I install Command Code?
+A: Run npm i -g command-code, then cmd login, then cmd to start.
+
+Q: What are Agent Skills?
+A: Reusable instructions that teach Command Code how to handle specific tasks like code reviews, testing, or commit messages. Use /skills to browse them.
+
+Q: What are MCP servers?
+A: MCP (Model Context Protocol) servers let Command Code connect to external tools and data sources like GitHub, Notion, databases, and more. Use /mcp to manage them.
+
+Q: What is plan mode?
+A: Plan mode lets you review and approve Command Code's approach before it writes any code. Press Shift+Tab to toggle it.
+
+Q: What is headless mode?
+A: Headless mode lets you run Command Code non-interactively in scripts, CI/CD pipelines, and automation workflows. Use cmd -p "query" to run.
+
+Q: How do I fix "command not found" after installing?
+A: Ensure your global npm bin directory is in your PATH. Run npm config get prefix to find it, then add <prefix>/bin to your PATH.
+
+Q: Why am I getting unauthorized errors?
+A: Run cmd logout followed by cmd login to refresh your session. If the issue persists, check your account at https://commandcode.ai/studio.
+
+Q: How do I check my usage limits?
+A: Visit the billing page in Command Code Studio at https://commandcode.ai/studio.
+
+Q: My MCP server won't connect. What should I do?
+A: Check the server config with cmd mcp get <name> and use /mcp inside a session to see the error. For stdio servers, ensure the command is installed and in your PATH.
+
+Q: Where can I get help?
+A: Join our Discord at https://commandcode.ai/discord, report issues on GitHub, or email support@commandcode.ai.
+
+## TROUBLESHOOTING
+
+### Installation
+- "command not found" after install: Ensure global npm bin directory is in PATH. Run npm config get prefix, add <prefix>/bin to PATH, restart terminal.
+
+### Authentication
+- Login fails or session expired: Run cmd logout then cmd login. If still failing, delete ~/.commandcode/auth.json and retry.
+- API key or provider issues: Verify provider config, check API key validity, check usage limits.
+- Chrome Local Network Access prompt during login: Click "Allow" — this is expected and safe. Command Code runs a temporary local server for OAuth callback.
+
+### MCP Servers
+- Server won't connect: Check config with cmd mcp get <name>, verify command is in PATH, use /mcp in session for error details.
+- Authentication issues: Re-authenticate with cmd mcp auth <server>, or clear and retry with cmd mcp auth --clear <server>.
+
+### Skills
+- Not appearing in /skills menu: Check directory structure (.commandcode/skills/ or .agents/skills/), verify SKILL.md exists with valid frontmatter.
+- Editor not opening: Check $EDITOR is set and the editor is in PATH.
+- Skill not being used: Make description more specific with keywords, ask Command Code explicitly to use the skill.
+
+### Performance
+- Slow responses: Check internet connection, disconnect unneeded MCP servers, keep prompts focused and specific.
+- High token usage: Review usage in Studio, use headless mode for scripted tasks, use plan mode to agree on approach first.
+
+### Support
+- Discord: https://commandcode.ai/discord
+- GitHub: https://github.com/commandcode/command-code
+- Email: support@commandcode.ai
+
+## PRICING & LIMITS
+How it works:
+- Monthly reset: Request limits reset at the start of each billing cycle.
+- Auto top-up: Buy extra credits at model cost. Credits roll over and never expire.
+- No AI training: Command Code does not train on your code or store your code snippets.
+
+Manage your plan:
+- Purchase credits or subscribe: Studio > Billing at https://commandcode.ai/billing
+- Track usage: https://commandcode.ai/usage (personal). For an org: https://commandcode.ai/<orgLogin>/settings/usage
+- Manage payment methods: Access billing portal from Studio > Billing
+
+Premium requests: Any request that uses a premium model (like taste-1) during your coding session.
+When you hit your limit: Continue with auto top-up enabled. Extra credits are purchased at model cost and roll over.
+Switch plans: Upgrade or downgrade anytime from Studio > Billing. Changes take effect next billing cycle.
+Team billing: Billed per seat. Credits are pooled at the team level.
+
+Enterprise: Enhanced security, compliance, and control. Contact support@commandcode.ai for Enterprise edition.
+
+## PRIVACY & LEGAL
+- Command Code does not train on your code or store your code snippets.
+- Taste processing runs on your codebase and stores learning data in your project and on your local machine only.
+- Privacy Policy: https://commandcode.ai/privacy
+- Terms of Service: https://commandcode.ai/terms
+
+## HELPFUL LINKS
+- Documentation: https://commandcode.ai/docs
+- Discord Community: https://commandcode.ai/discord