Initial release: Edition 1.3 — 28 tools, 3 platforms, fully verified

Claude's Hidden Toolkit documents 28 undocumented internal tools across
Claude.ai's browser, desktop app, and mobile app interfaces. All parameter
schemas and response formats are empirically confirmed through cross-platform
testing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.github/ISSUE_TEMPLATE/correction.yml

@@ -0,0 +1,47 @@
+name: "🐛 Correction"
+description: Report an error or outdated information in the book
+title: "[Correction] "
+labels: ["correction"]
+body:
+  - type: input
+    id: location
+    attributes:
+      label: Location in Book
+      description: Which chapter, tool card, or section?
+      placeholder: "e.g., Tool Card 5 (message_compose_v1), Chapter 7"
+    validations:
+      required: true
+  - type: textarea
+    id: current
+    attributes:
+      label: What the Book Currently Says
+      description: Quote or paraphrase the incorrect/outdated content
+    validations:
+      required: true
+  - type: textarea
+    id: observed
+    attributes:
+      label: What You Observed Instead
+      description: Describe the correct behavior with evidence
+    validations:
+      required: true
+  - type: dropdown
+    id: platform
+    attributes:
+      label: Platform Tested On
+      multiple: true
+      options:
+        - Browser (claude.ai)
+        - Desktop App (macOS)
+        - Desktop App (Windows)
+        - Mobile App (Android)
+        - Mobile App (iOS)
+    validations:
+      required: true
+  - type: input
+    id: date
+    attributes:
+      label: Date of Observation
+      placeholder: "2026-02-16"
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/discovery.yml

@@ -0,0 +1,65 @@
+name: "🔍 New Tool Discovery"
+description: Report a newly discovered tool or capability
+title: "[Discovery] "
+labels: ["discovery"]
+body:
+  - type: input
+    id: tool-name
+    attributes:
+      label: Tool Name
+      description: The internal tool name (e.g., `reminder_create_v0`)
+      placeholder: tool_name_v0
+    validations:
+      required: true
+  - type: dropdown
+    id: platform
+    attributes:
+      label: Platform
+      description: Where did you discover this?
+      multiple: true
+      options:
+        - Browser (claude.ai)
+        - Desktop App (macOS)
+        - Desktop App (Windows)
+        - Mobile App (Android)
+        - Mobile App (iOS)
+    validations:
+      required: true
+  - type: dropdown
+    id: discovery-method
+    attributes:
+      label: Discovery Method
+      options:
+        - tool_search query
+        - System prompt extraction
+        - Behavioral observation
+        - Error message / stack trace
+        - Other
+    validations:
+      required: true
+  - type: textarea
+    id: schema
+    attributes:
+      label: Parameter Schema
+      description: JSON schema if available (from tool_search results or system prompt)
+      render: json
+  - type: textarea
+    id: response
+    attributes:
+      label: Response Example
+      description: A real response from calling the tool
+      render: json
+  - type: textarea
+    id: behavior
+    attributes:
+      label: Observed Behavior
+      description: What does the tool do? Any gotchas, edge cases, or platform differences?
+    validations:
+      required: true
+  - type: input
+    id: date
+    attributes:
+      label: Date of Discovery
+      placeholder: "2026-02-16"
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/screenshot.yml

@@ -0,0 +1,48 @@
+name: "📸 Screenshot Contribution"
+description: Submit a screenshot for one of the 22 [Upcoming] figures
+title: "[Screenshot] fig_XX — "
+labels: ["screenshot"]
+body:
+  - type: input
+    id: figure
+    attributes:
+      label: Figure Number
+      description: Which figure are you providing? (fig_01 through fig_22)
+      placeholder: "fig_07"
+    validations:
+      required: true
+  - type: dropdown
+    id: platform
+    attributes:
+      label: Platform Captured On
+      options:
+        - Browser (claude.ai)
+        - Desktop App (macOS)
+        - Desktop App (Windows)
+        - Mobile App (Android)
+        - Mobile App (iOS)
+    validations:
+      required: true
+  - type: textarea
+    id: description
+    attributes:
+      label: What's Shown
+      description: Brief description of what the screenshot captures
+    validations:
+      required: true
+  - type: textarea
+    id: screenshot
+    attributes:
+      label: Screenshot
+      description: Drag and drop your screenshot here. Please blur any personal information.
+    validations:
+      required: true
+  - type: checkboxes
+    id: privacy
+    attributes:
+      label: Privacy Check
+      options:
+        - label: I have blurred or removed all personal information from this screenshot
+          required: true
+        - label: I consent to this image being used in the book under CC BY 4.0
+          required: true

.gitignore

@@ -0,0 +1,25 @@
+# OS
+.DS_Store
+Thumbs.db
+desktop.ini
+
+# Editors
+*.swp
+*.swo
+*~
+.vscode/
+.idea/
+
+# Temp files
+*.tmp
+*.bak
+*.orig
+
+# Build artifacts
+__pycache__/
+node_modules/
+*.pyc
+
+# Draft versions (keep only release versions in repo)
+*_draft*
+*_WIP*

AGENTS.md

@@ -0,0 +1,46 @@
+# AGENTS.md — Context for AI Coding Agents
+
+## What This Repo Is
+
+A technical book ("Claude's Hidden Toolkit") documenting 28 undocumented internal tools in Claude.ai's consumer interfaces. Fully reverse-engineered through systematic testing across browser, desktop app, and mobile app.
+
+## File Roles
+
+| File | Role | Editable? |
+|------|------|-----------|
+| `Claude_Hidden_Toolkit.md` | Source of truth | Yes — all content changes here |
+| `README.md` | GitHub landing page | Yes |
+| `LICENSE` | CC BY 4.0 | No |
+
+## Domain Knowledge
+
+- **28 tools** documented with confirmed JSON schemas
+- **3 platforms** with different tool inventories (browser: 21 tools, desktop: 22+32 MCP, mobile: 20+11 deferred)
+- **Tool cards** in the Technical Annex follow a strict template (see Appendix C in the book)
+- **Figures** are numbered sequentially fig_01 through fig_22 — all currently marked [Upcoming] pending screenshot capture
+- **Response formats** are all empirically confirmed — no inferred markers remain
+
+## Conventions
+
+- Tool names in backticks: `web_search`, `event_create_v0`
+- Platform tags on every claim: "Mobile App only", "Browser + Desktop only", "All"
+- Version suffixes: `_v0` (initial/experimental), `_v1` (breaking upgrade)
+- Confirmed vs unverified clearly marked — never speculate
+
+## Architecture Context
+
+The book documents a two-tier tool loading system:
+- **Always-loaded tools**: injected into every conversation context (~17-22 depending on platform)
+- **Deferred tools**: discovered at runtime via `tool_search` (mobile only, 11 consumer tools)
+- **MCP tools**: discovered via `tool_search` on desktop only (32 integration tools)
+
+The calendar system has a dual-layer architecture where Projects use MCP Connector tools with email-based IDs, while standard mobile chats use deferred consumer tools with numeric string IDs. The two layers are incompatible.
+
+## Quality Gates
+
+Before any content change:
+1. Has the claim been tested on the relevant platform(s)?
+2. Is the platform availability clearly stated?
+3. Does the tool card follow the template structure?
+4. Are figure numbers sequential with no gaps?
+5. Is the response format marked as confirmed or unverified?

CHANGELOG.md

@@ -0,0 +1,68 @@
+# Changelog
+
+All notable changes to Claude's Hidden Toolkit are documented here.
+
+Format follows [Keep a Changelog](https://keepachangelog.com/). Editions are tagged by version.
+
+## [1.3] — 2026-02-16
+
+### Added
+- **Platform Architecture section** (new Chapter 2.5) — tool inventory matrix across browser (21 tools), desktop app (22+32 MCP), and mobile app (20+11 deferred)
+- **3 new tool cards**: `weather_fetch` (Card 26), `recipe_display_v0` (Card 27), `end_conversation` (Card 28)
+- Platform-exclusive tools table showing which tools exist where
+- Session instability warning (switching clients changes available tools)
+- Fallback behavior documentation for each platform gap
+- Settings & Connectors reference integrated into appendix
+- 9 new figure placeholders (fig_14 through fig_22)
+
+### Changed
+- **Chapter 10 (Discovery Log)**: added mobile-only environment note — the discovery log was produced on mobile and is not reproducible on browser or desktop
+- **Chapter 2 (Architecture)**: rewritten to reflect three platform configurations instead of a single binary always-loaded/deferred model
+- **Card 5 (message_compose_v1)**: corrected primary button to "Send via Gmail", documented dropdown with "Open in Mail", added no-recipient-field note
+- **Card 4 (ask_user_input_v0)**: added "Something else" free-text option, Skip button, keyboard shortcuts, selection counter, drag handles, pagination
+- **Cards 17/18 (places)**: added markers mode sidebar details, itinerary mode routing, "Open route" Google Maps behavior, confirmation dialog inconsistency
+- **Card 6 (alarm_create_v0)**: confirmed mobile-only, added widget description with Italian locale, removed web/desktop speculation
+- **Card 7 (timer_create_v0)**: confirmed mobile-only, added widget description
+- **Card 1 (user_time_v0)**: added platform behavior differences (mobile returns exact time, desktop/browser states system date only)
+- **Card 3 (user_location_v0)**: confirmed mobile-only, documented two-layer permission system (Claude app + Android OS)
+- **Card 10 (chart_display_v0)**: confirmed mobile-only deferred, added app crash warning
+- **Card 25 (tool_search)**: major rewrite — browser has no tool_search, desktop finds MCP only, mobile finds consumer tools
+- **Cards 8–13 (Calendar suite)**: added mobile-only deferred status, Google Calendar connection requirement
+- **Appendix A**: corrected availability matrix to reflect cross-platform testing results
+- All response formats upgraded from *(inferred)* to empirically confirmed
+- Figure numbering made sequential (fig_01 through fig_22, eliminated fig_14b)
+
+### Removed
+- All *(inferred)* response format markers — every schema now confirmed through testing
+- Speculative web/desktop availability claims for mobile-only tools
+- Screenshot capture instructions from figure placeholders
+
+## [1.2] — 2026-02-13
+
+### Added
+- Technical Annex with 25 detailed tool cards
+- Appendix C: Tool Template for New Discoveries
+- Figure placeholders with capture instructions
+- Companion GitHub repository
+
+### Changed
+- Expanded Chapter 10 Discovery Log with full query-by-query enumeration
+- Added environment note about tool_search behavior in Projects vs standard chats
+
+## [1.1] — 2026-02-10
+
+### Added
+- Complete parameter schemas for all 25 tools
+- Cross-reference with prior research (Khemani, Willison, Adversa AI, Viticci)
+- Discovery methodology (Appendix B)
+
+### Changed
+- Restructured from flat tool list to chapter-based organization
+
+## [1.0] — 2026-02-07
+
+### Added
+- Initial release documenting 25 internal tools
+- Two-tier architecture analysis (always-loaded vs deferred)
+- tool_search enumeration methodology
+- Availability matrix (Appendix A)

CITATION.cff

@@ -0,0 +1,27 @@
+cff-version: 1.2.0
+message: "If you reference this work, please cite it as below."
+title: "Claude's Hidden Toolkit: The Definitive Guide to Claude.ai's Undocumented Internal Tools"
+type: software
+authors:
+  - family-names: "Marinoni Moretto"
+    given-names: "Francesco"
+    orcid: ""
+version: "1.3"
+date-released: "2026-02-16"
+url: "https://github.com/n1-ai/claude-hidden-toolkit"
+license: "CC-BY-4.0"
+keywords:
+  - claude
+  - anthropic
+  - ai-tools
+  - reverse-engineering
+  - internal-tools
+  - tool-use
+  - model-context-protocol
+  - claude-ai
+abstract: >-
+  Comprehensive reverse-engineering documentation of 28 undocumented internal
+  tools in Claude.ai's consumer interfaces (browser, desktop app, mobile app).
+  Covers complete parameter schemas, confirmed response formats, platform-specific
+  behavior differences, and a three-tier tool loading architecture across three
+  client surfaces. Based on 35+ manual tests with empirical verification.

CLAUDE.md

@@ -0,0 +1,49 @@
+# CLAUDE.md — Instructions for Claude Code
+
+## Project Overview
+
+This is "Claude's Hidden Toolkit" — a living technical book documenting Claude.ai's undocumented internal tools. The book reverse-engineers 28 tools across three platforms (browser, desktop app, mobile app) with confirmed schemas, response formats, and behavioral patterns.
+
+## Repository Structure
+
+```
+Claude_Hidden_Toolkit.md   # The book (source of truth — all edits here)
+README.md                  # GitHub landing page
+CLAUDE.md                  # You are here
+AGENTS.md                  # General AI agent context
+CHANGELOG.md               # Edition history with detailed changes
+CITATION.cff               # Academic citation metadata
+CONTRIBUTING.md            # Contribution guidelines
+SECURITY.md                # Responsible disclosure policy
+LICENSE                    # CC BY 4.0
+```
+
+## Key Rules
+
+1. **The .md file is the source of truth.** All content edits happen here.
+2. **All tool claims must be empirically verified.** No speculative schemas. If you haven't tested it, mark it as unverified.
+3. **Platform matters.** Every tool claim must specify which platform(s) it applies to: browser, desktop app, mobile app, or all.
+4. **Figure numbering is sequential** (fig_01 through fig_22). If adding figures, continue the sequence. Never use letter suffixes (e.g., fig_14b).
+5. **Tool cards follow a strict template.** See Appendix C in the book for the template. Every card needs: quick facts table, overview, parameters table, example call with JSON, "When Claude Uses It" table, "Watch Out" section, and platform availability table.
+
+## Writing Style
+
+- Technical but accessible — assume the reader is a developer or power user
+- No hedging without evidence — say "confirmed" or "not tested", never "probably"
+- Use backtick formatting for all tool names, parameter names, and code
+- Tables over bullet lists for structured comparisons
+- Blockquotes (>) for figure placeholders and warnings
+
+## Tool Naming Conventions
+
+- Always-loaded tools: `web_search`, `memory_user_edits`, etc.
+- Deferred tools (mobile only): `user_time_v0`, `alarm_create_v0`, etc.
+- Version suffixes: `_v0` = initial, `_v1` = breaking change upgrade
+- Unversioned tools are considered stable
+
+## What Not To Do
+
+- Don't add tools without cross-platform verification
+- Don't change the figure numbering scheme without renumbering all references
+- Don't remove the [Upcoming] tags on figures — they indicate screenshots pending capture
+- Don't merge the .md living repository links with personal GitHub links — they serve different purposes

CONTRIBUTING.md

@@ -0,0 +1,76 @@
+# Contributing to Claude's Hidden Toolkit
+
+Thanks for your interest in contributing! This project documents Claude.ai's undocumented internal tools — every contribution helps the community understand what's really happening under the hood.
+
+## How to Contribute
+
+### Reporting New Tools or Changes
+
+If you've discovered a new tool, a schema change, or platform-specific behavior not covered in the book:
+
+1. **Open an Issue** with the label `discovery`
+2. Include:
+   - Tool name (exact internal name if known)
+   - Platform tested on (browser / desktop app / mobile app / iOS / Android)
+   - How you discovered it (e.g., `tool_search` query, system prompt extraction, behavioral observation)
+   - Parameter schema if available (JSON)
+   - Response format with a real example
+   - Any gotchas or edge cases observed
+
+### Correcting Existing Documentation
+
+If something in the book is wrong or outdated:
+
+1. **Open an Issue** with the label `correction`
+2. Include:
+   - Which tool card or chapter section is affected
+   - What the book currently says
+   - What you observed instead
+   - Platform and date of observation
+
+### Submitting a Pull Request
+
+For direct content changes:
+
+1. Fork the repo
+2. Edit `Claude_Hidden_Toolkit.md` only — this is the source of truth
+3. Follow the existing formatting conventions (see below)
+4. Submit a PR with a clear description of what changed and how it was verified
+
+## Formatting Conventions
+
+- **Tool names** in backticks: `web_search`, `event_create_v0`
+- **Platform availability** stated on every claim
+- **Tool cards** follow the template in Appendix C — don't freestyle the structure
+- **Figures** are numbered sequentially (fig_01 through fig_22+). Continue the sequence; never reuse or skip numbers
+- **Response formats** marked as confirmed with real output. If you can't confirm, note it explicitly
+- **Tables** for structured comparisons, not bullet lists
+
+## Verification Standard
+
+This project has a strict empirical standard. We don't document tools based on speculation, hearsay, or what "should" work. Every claim must be:
+
+- **Tested** on at least one platform with a real Claude.ai session
+- **Reproducible** — another user should be able to trigger the same behavior
+- **Platform-tagged** — specifying exactly where it works and where it doesn't
+
+If you've observed something but can't fully verify it, open an Issue rather than a PR. The community can help confirm.
+
+## What We're Looking For
+
+High-value contributions include:
+
+- **iOS-specific tools** (Reminders suite, Siri integration) — we have limited iOS testing coverage
+- **Screenshot captures** for the 22 figures currently marked [Upcoming]
+- **Cross-platform verification** of existing tool cards on platforms we haven't tested
+- **New deferred tools** discovered via `tool_search` sweeps
+- **Schema changes** from Anthropic updates that break or modify existing tools
+- **MCP Connector behavior** differences between Projects and standard chats
+
+## Code of Conduct
+
+Be respectful, constructive, and evidence-based. This is a technical documentation project — claims need proof, disagreements are settled by testing, and everyone's contributions are valued.
+
+## License
+
+By contributing, you agree that your contributions will be licensed under [CC BY 4.0](LICENSE), the same license as the rest of the project.

LICENSE

@@ -0,0 +1,24 @@
+Creative Commons Attribution 4.0 International License (CC BY 4.0)
+
+Copyright (c) 2026 Francesco Marinoni Moretto
+
+You are free to:
+
+  Share — copy and redistribute the material in any medium or format for
+  any purpose, even commercially.
+
+  Adapt — remix, transform, and build upon the material for any purpose,
+  even commercially.
+
+Under the following terms:
+
+  Attribution — You must give appropriate credit, provide a link to the
+  license, and indicate if changes were made. You may do so in any
+  reasonable manner, but not in any way that suggests the licensor
+  endorses you or your use.
+
+  No additional restrictions — You may not apply legal terms or
+  technological measures that legally restrict others from doing
+  anything the license permits.
+
+Full license text: https://creativecommons.org/licenses/by/4.0/legalcode

README.md

@@ -0,0 +1,79 @@
+# Claude's Hidden Toolkit
+
+**The Definitive Guide to Claude.ai's Undocumented Internal Tools**
+
+*28 tools. Zero official documentation. Fully reverse-engineered.*
+
+---
+
+## What Is This?
+
+Claude.ai ships with 28+ internal tools — capabilities that let Claude check the time, search your calendar, render interactive charts, draft emails with native app integration, find nearby restaurants, display them on a map, set alarms on your phone, and more.
+
+**Almost none of this is documented.**
+
+Anthropic's official docs cover the API. But the consumer product — claude.ai — has a rich, evolving toolkit that exists in a documentation void. This book fills that gap through systematic reverse-engineering across three platforms (browser, desktop app, mobile app).
+
+## Key Discoveries
+
+**Platform fragmentation is real.** Claude is not one product — it's three different tool sets:
+
+| Platform | Always-Loaded | tool_search | Deferred Tools |
+|----------|:---:|:---:|:---:|
+| Browser (claude.ai) | 21 | ❌ | None |
+| Desktop App | 22 | MCP only | 32 (Chrome + Filesystem) |
+| Mobile App | 20 | Consumer | 11 (alarm, timer, calendar, etc.) |
+
+**Every response format has been empirically confirmed** — no inferred schemas remain in Edition 1.3.
+
+## What's Inside
+
+| Chapter | Coverage |
+|---------|----------|
+| 1–2 | Architecture: two-tier loading, platform inventory, versioning |
+| 3 | Context tools: `user_time_v0`, `user_location_v0` |
+| 4 | Interaction widgets: `ask_user_input_v0`, `message_compose_v1` |
+| 5 | Inline charts: `chart_display_v0` |
+| 6 | Calendar & device: 6 calendar tools + alarms + timers |
+| 7 | Search & data: web, images, places, maps, sports, Drive, weather, recipes |
+| 8 | Memory: `memory_user_edits`, `conversation_search`, `recent_chats` |
+| 9 | Computer use: bash, files, artifacts, the skill system |
+| 10 | The meta-tool: `tool_search` and enumeration methodology |
+| 11 | Complete tool reference index |
+| 12 | MCP convergence and what comes next |
+| Annex | 28 detailed tool cards with confirmed JSON schemas |
+
+## Read the Book
+
+- **[Claude_Hidden_Toolkit.md](Claude_Hidden_Toolkit.md)** — Full book, Markdown (~21,600 words)
+
+## Edition History
+
+| Edition | Date | Changes |
+|---------|------|---------|
+| 1.3 | Feb 2026 | Platform architecture chapter, 3 new tool cards (weather, recipe, end_conversation), cross-platform verification of all 28 cards, settings/connectors appendix, all response formats confirmed |
+| 1.2 | Feb 2026 | 25 tool cards, discovery methodology, initial availability matrix |
+| 1.0–1.1 | Feb 2026 | Initial discovery and documentation |
+
+## Prior Art & Credits
+
+This work builds on contributions from: **Shlok Khemani** (memory tool schemas, Sept 2025), **Simon Willison** (system prompt analysis), **Adversa AI** (security research, named `message_compose_v1`), **Federico Viticci / MacStories** (iOS tools and `_v0` naming convention discovery).
+
+## Who This Is For
+
+- **Power users** — learn exactly what to say to trigger hidden capabilities
+- **Developers** — see what the consumer product does that the API doesn't
+- **AI researchers** — a documented case study of production AI tool architecture
+
+## Author
+
+**Francesco Marinoni Moretto** — AI Practice & Community Lead, [N1AI](https://n1.ai)
+
+- LinkedIn: [linkedin.com/in/francesco-moretto](https://www.linkedin.com/in/francesco-moretto)
+- GitHub: [github.com/frmoretto](https://github.com/frmoretto)
+
+## License
+
+CC BY 4.0 — You are free to share and adapt this work with attribution.
+
+This work is independent and not affiliated with Anthropic. Tools may change, be renamed, or be removed at any time.

SECURITY.md

@@ -0,0 +1,46 @@
+# Security Policy
+
+## Scope
+
+This project documents Claude.ai's internal tools through reverse-engineering. While the book itself is documentation (not executable code), the subject matter touches on AI system internals, and discoveries may occasionally surface behaviors with security implications.
+
+## Reporting a Vulnerability
+
+If you discover a tool behavior that could be exploited to:
+
+- Bypass Claude's safety guardrails
+- Access other users' data through shared storage or memory tools
+- Exfiltrate sensitive information via tool parameters or responses
+- Manipulate tool behavior through prompt injection in tool inputs
+- Escalate permissions beyond intended scope
+
+**Please report it responsibly:**
+
+1. **Do NOT open a public Issue.** Security-relevant findings should not be disclosed publicly before they can be addressed.
+2. **Email:** Contact the maintainer directly at the email listed on the [author's GitHub profile](https://github.com/frmoretto).
+3. **Include:**
+   - Description of the behavior
+   - Steps to reproduce
+   - Platform(s) affected
+   - Potential impact assessment
+4. **Also consider reporting to Anthropic directly** via their [responsible disclosure process](https://www.anthropic.com/responsible-disclosure-policy) — they are best positioned to fix tool-level vulnerabilities.
+
+## What This Project Will Do
+
+- Acknowledge receipt within 48 hours
+- Assess whether the finding should be withheld from publication until Anthropic addresses it
+- Coordinate with Anthropic if appropriate
+- Credit the reporter (unless anonymity is requested) when the finding is eventually published
+
+## What This Project Will NOT Do
+
+- Publish exploit details before Anthropic has had reasonable time to respond
+- Document tools or behaviors specifically designed to facilitate harm
+- Provide instructions for bypassing safety systems
+
+## Supported Versions
+
+| Version | Supported |
+|---------|-----------|
+| 1.3     | Yes       |
+| < 1.3   | No — please reference the latest edition |