Skip to content
Claude Code Docs Claude Code Docs

Publicly Documented Bundled Skills

Publicly-documented bundled skills

Section titled “Publicly-documented bundled skills”

Code operations

Section titled “Code operations”

/batch

Section titled “/batch”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/batch.ts
  • Aliases: none
  • User-invocable: yes (required)
  • Model-invocable: yes
  • Gating: always enabled
  • Argument hint: <instruction>
  • Allowed tools: inherits from session (tool use via agents)
  • Model override: none
  • Description: Research and plan a large-scale change, then execute it in parallel across 5–30 isolated worktree agents that each open a PR.
  • When to use: Use when the user wants to make a sweeping, mechanical change across many files (migrations, refactors, bulk renames) that can be decomposed into independent parallel units. Example: /batch migrate from React to Vue.
  • Example: /batch migrate src/ from Solid to React
  • Notes: Requires a git repository. Spawns independent agents in isolated worktrees (via isolation: "worktree"). Requires git and GitHub CLI (gh). Phase 1 is plan mode where work is decomposed; Phase 2 spawns all worker agents in parallel; Phase 3 tracks completion and PR links.

/simplify

Section titled “/simplify”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/simplify.ts
  • Aliases: none
  • User-invocable: yes (default)
  • Model-invocable: yes
  • Gating: always enabled
  • Argument hint: [focus] (optional)
  • Allowed tools: inherits from session
  • Model override: none
  • Description: Review changed code for reuse, quality, and efficiency, then fix any issues found.
  • When to use: Use after making code changes to audit for redundancy, anti-patterns, and performance issues. Spawns three concurrent review agents (code reuse, quality, efficiency) and aggregates findings.
  • Example: /simplify focus on memory efficiency
  • Notes: Reads git diff to identify changed files. Non-git workflows fall back to recently modified files. Agent-based; runs in parallel via Agent tool.

/debug

Section titled “/debug”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/debug.ts
  • Aliases: none
  • User-invocable: yes (required)
  • Model-invocable: no (disableModelInvocation: true)
  • Gating: always enabled
  • Argument hint: [issue description] (optional)
  • Allowed tools: Read, Grep, Glob
  • Model override: none
  • Description: Enable debug logging for this session and help diagnose issues.
  • When to use: Use when debugging a problem in the current Claude Code session. Debug logging is off by default unless started with claude --debug; running /debug mid-session enables logging from that point forward.
  • Example: /debug Claude is hanging during file reads or just /debug
  • Notes: For ant users, description shows “Includes all event logging”. Non-ant users see shorter description. Tails the last 64KB of the debug log and reads last 20 lines. Launches the “Claude Code Guide” subagent for feature explanations if needed.

Session & conversation

Section titled “Session & conversation”

/loop

Section titled “/loop”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/loop.ts
  • Aliases: /proactive
  • User-invocable: yes (default)
  • Model-invocable: yes
  • Gating: feature-flag gated (isKairosCronEnabled() — must have Kairos cron backend available)
  • Argument hint: [interval] <prompt>
  • Allowed tools: inherits from session (uses Cron Create/Delete tools internally)
  • Model override: none
  • Description: Run a prompt or slash command on a recurring interval (e.g. /loop 5m /foo, defaults to 10m).
  • When to use: When the user wants to set up a recurring task, poll for status, or run something repeatedly on an interval (e.g. “check the deploy every 5 minutes”, “keep running /babysit-prs”). Do NOT invoke for one-off tasks.
  • Example: /loop 5m check if the deploy finished or /loop 30m /babysit-prs
  • Notes: Parses interval from leading token (5m, 2h, 1d) or trailing every clause. Defaults to 10m if no interval given. Converts intervals to cron expressions. Minimum granularity is 1 minute; seconds are rounded up. Tasks auto-expire after 30 days. Invokes the prompt immediately after scheduling (does not wait for first cron fire).

/stuck (internal)

Section titled “/stuck (internal)”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/stuck.ts
  • Aliases: none
  • User-invocable: yes
  • Model-invocable: no (disableModelInvocation: true)
  • Gating: ant-only (USER_TYPE !== 'ant' → not registered)
  • Argument hint: [pid or symptom] (optional)
  • Allowed tools: inherits from session (for bash diagnostics)
  • Model override: none
  • Description: [ANT-ONLY] Investigate frozen/stuck/slow Claude Code sessions on this machine and post a diagnostic report to #claude-code-feedback.
  • When to use: Use when another Claude Code session on the same machine appears frozen, stuck, or very slow. Diagnosis includes CPU, memory, process state, child processes, and optional stack dumps.
  • Notes: Scans ps output for hung processes (state D = uninterruptible sleep, state T = stopped, state Z = zombie). Looks for high CPU (≥90%) or high RSS (≥4GB). Posts diagnostic report to Slack #claude-code-feedback via MCP if available. Ant staff only.

Integrations & API

Section titled “Integrations & API”

/claude-api

Section titled “/claude-api”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/claudeApi.ts
  • Aliases: none
  • User-invocable: yes (default)
  • Model-invocable: yes (auto-triggers when code imports anthropic or @anthropic-ai/sdk)
  • Gating: always enabled
  • Argument hint: none
  • Allowed tools: Read, Grep, Glob, WebFetch
  • Model override: none
  • Description: Build apps with the Claude API or Anthropic SDK. Loads language-specific Claude API reference material (Python, TypeScript, Java, Go, Ruby, C#, PHP, cURL). Covers tool use, streaming, batches, structured outputs, Managed Agents, and common pitfalls.
  • When to use: Use when code imports anthropic/@anthropic-ai/sdk/claude_agent_sdk, or user asks to use Claude API, Anthropic SDKs, or Agent SDK. Do NOT trigger when code imports openai/other AI SDK, general programming, or ML/data-science tasks.
  • Reference files extracted: Yes — 247KB of language-specific documentation bundled and lazy-loaded. Extracted on first invocation to enable Read/Grep queries on reference docs.
  • Example: /claude-api How do I stream responses? or auto-invoked when you import the SDK
  • Notes: Auto-detects project language by scanning for language indicators (.py files, package.json, go.mod, etc.). If no language detected, includes all docs and asks user. Reference material includes README, streaming, tool-use, batches, files-API, prompt-caching, error codes, and live links to web docs.

/claude-in-chrome

Section titled “/claude-in-chrome”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/claudeInChrome.ts
  • Aliases: none
  • User-invocable: yes (default)
  • Model-invocable: yes (can auto-trigger when user has Claude in Chrome extension configured)
  • Gating: shouldAutoEnableClaudeInChrome() — checks if extension is installed and configured
  • Argument hint: none
  • Allowed tools: MCP tools from Claude in Chrome extension (browser automation, screenshots, console logs, etc.)
  • Model override: none
  • Description: Automates your Chrome browser to interact with web pages - clicking elements, filling forms, capturing screenshots, reading console logs, and navigating sites. Opens pages in new tabs within your existing Chrome session. Requires site-level permissions before executing (configured in the extension).
  • When to use: When the user wants to interact with web pages, automate browser tasks, capture screenshots, read console logs, or perform any browser-based actions. Always invoke BEFORE attempting to use any mcp__claude-in-chrome__* tools.
  • Example: /claude-in-chrome Take a screenshot of the login form
  • Notes: Wraps MCP tools from the Claude in Chrome browser extension. First call mcp__claude-in-chrome__tabs_context_mcp to learn about current tabs. Gating checks shouldAutoEnableClaudeInChrome() which verifies the extension is installed.

Configuration & setup

Section titled “Configuration & setup”

/update-config

Section titled “/update-config”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/updateConfig.ts
  • Aliases: none
  • User-invocable: yes (default)
  • Model-invocable: yes
  • Gating: always enabled
  • Argument hint: none
  • Allowed tools: Read (for reading existing settings files)
  • Model override: none
  • Description: Configure Claude Code via settings.json. Use for hooks (“run X before/after Y”), permissions (“allow Z”), environment variables, MCP configuration, and any settings.json modifications.
  • When to use: Use when the user wants to automate behaviors via hooks, add permissions, set environment variables, or modify any .claude/settings.json or .claude/settings.local.json file. For simple settings like theme/model, the Config tool is preferred.
  • Example: /update-config allow npm commands without prompting or /update-config set DEBUG=true
  • Notes: Includes comprehensive hook documentation with examples (PreToolUse, PostToolUse, PreCompact, etc.). Always reads existing file before writing (merge-first strategy). Provides hook verification workflow: construct, pipe-test, validate, prove, deploy. Distinguishes when hooks are required vs. memory is sufficient.

/keybindings-help

Section titled “/keybindings-help”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/keybindings.ts
  • Aliases: none
  • User-invocable: no (user-invocable: false)
  • Model-invocable: yes (only Claude invokes)
  • Gating: isKeybindingCustomizationEnabled() — checks if feature is available
  • Argument hint: none
  • Allowed tools: Read
  • Model override: none
  • Description: Use when the user wants to customize keyboard shortcuts, rebind keys, add chord bindings, or modify ~/.claude/keybindings.json.
  • When to use: Use for keybinding customization requests like “rebind ctrl+s”, “add a chord shortcut”, “change the submit key”.
  • Reference files extracted: no
  • Example: Claude may invoke when you ask “How do I rebind the save key?”
  • Notes: Generates reference tables dynamically from source-of-truth arrays (contexts, actions, reserved shortcuts). Includes keystroke syntax guide (modifiers: ctrl, alt, shift, meta; chords with 1-second timeout between). Validates against reserved shortcuts (terminal, macOS). /doctor command validates keybinding syntax.

/schedule

Section titled “/schedule”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/scheduleRemoteAgents.ts
  • Aliases: none
  • User-invocable: yes (default)
  • Model-invocable: yes
  • Gating: feature-flag (tengu_surreal_dali) + policy check (allow_remote_sessions)
  • Argument hint: none
  • Allowed tools: RemoteTrigger (for API calls), AskUserQuestion
  • Model override: none
  • Description: Create, update, list, or run scheduled remote agents (triggers) that execute on a cron schedule in Anthropic’s cloud infrastructure.
  • When to use: When the user wants to schedule a recurring remote agent, set up automated tasks, create a cron job for Claude Code, or manage their scheduled agents/triggers.
  • Example: /schedule every morning at 9am check the deploy status
  • Notes: Requires claude.ai account authentication (API accounts not supported). Minimum cron interval is 1 hour. Converts user times to UTC using their local timezone. Supports attaching MCP connectors (Slack, Datadog, etc.). Creates default environment if none exist. Cross-checks GitHub app permissions if triggering repo changes. Remote agents run in isolated cloud environments with no access to local machine/files.

Developer / internal

Section titled “Developer / internal”

/verify (internal)

Section titled “/verify (internal)”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/verify.ts
  • Aliases: none
  • User-invocable: yes
  • Model-invocable: no
  • Gating: ant-only (USER_TYPE !== 'ant' → not registered)
  • Argument hint: none
  • Allowed tools: inherits from session
  • Model override: none
  • Description: Verify a code change does what it should by running the app.
  • When to use: Use for testing and verifying that code changes work as intended.
  • Reference files extracted: yes (via files: SKILL_FILES from verifyContent.ts)
  • Notes: Ant-only skill. Content and reference files imported from verifyContent.ts. Allows optional user request argument for focused verification.

/remember (internal)

Section titled “/remember (internal)”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/remember.ts
  • Aliases: none
  • User-invocable: yes
  • Model-invocable: no
  • Gating: auto-memory enabled (isAutoMemoryEnabled())
  • Argument hint: none
  • Allowed tools: inherits from session (Read, Edit, etc.)
  • Model override: none
  • Description: Review auto-memory entries and propose promotions to CLAUDE.md, CLAUDE.local.md, or shared memory. Also detects outdated, conflicting, and duplicate entries across memory layers.
  • When to use: Use when the user wants to review, organize, or promote their auto-memory entries. Also useful for cleaning up outdated or conflicting entries across CLAUDE.md, CLAUDE.local.md, and auto-memory.
  • Notes: Only registered if isAutoMemoryEnabled() returns true. Classifies auto-memory entries into destinations: CLAUDE.md (team conventions), CLAUDE.local.md (personal preferences), team memory (org-wide). Detects duplicates and conflicts. Proposes changes but does not modify without explicit approval.

/lorem-ipsum (internal)

Section titled “/lorem-ipsum (internal)”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/loremIpsum.ts
  • Aliases: none
  • User-invocable: yes
  • Model-invocable: no
  • Gating: ant-only (USER_TYPE !== 'ant' → not registered)
  • Argument hint: [token_count]
  • Allowed tools: none
  • Model override: none
  • Description: Generate filler text for long context testing. Specify token count as argument (e.g., /lorem-ipsum 50000). Outputs approximately the requested number of tokens.
  • Example: /lorem-ipsum 10000
  • Notes: Ant-only. Capped at 500,000 tokens for safety. Generates random sentences from a curated list of 200+ one-token English words. Paragraphs inserted randomly (~20% chance after each sentence).

/skillify (internal)

Section titled “/skillify (internal)”
  • Type: Bundled skill
  • Source: claude-code-main/skills/bundled/skillify.ts
  • Aliases: none
  • User-invocable: yes
  • Model-invocable: no (disableModelInvocation: true)
  • Gating: ant-only (USER_TYPE !== 'ant' → not registered)
  • Argument hint: [description of the process you want to capture] (optional)
  • Allowed tools: Read, Write, Edit, Glob, Grep, AskUserQuestion, Bash(mkdir:*)
  • Model override: none
  • Description: Capture this session’s repeatable process into a skill. Call at end of the process you want to capture with an optional description.
  • When to use: Use at the end of a complex workflow you want to automate and reuse. Example: After completing a cherry-pick workflow, run /skillify cherry-pick workflow to capture the steps.
  • Notes: Ant-only. Interviews user via AskUserQuestion in multiple rounds: high-level confirmation, detailed breakdown, per-step analysis, final questions. Reads session memory and user messages to understand the context. Generates SKILL.md in user’s chosen location (.claude/skills/ or ~/.claude/skills/). Includes frontmatter for metadata and detailed step-by-step instructions with success criteria and annotations.

← Back to Skills/README.md