A repository containing all system prompts, tool descriptions, and agent prompts extracted from Claude Code, updated with each release.
Claude Code System Prompts is a reference repository that extracts and archives all the internal system prompts, tool descriptions, and agent instructions used by Claude Code. It solves the problem of opacity in AI-powered coding tools by providing developers with a transparent, versioned view of the exact instructions that govern the assistant's behavior, from simple file edits to complex multi-agent planning workflows.
Developers, researchers, and prompt engineers who use or study Claude Code and want to understand its internal logic, track changes between versions, or create custom modifications using tools like `tweakcc`.
It offers unique, direct access to the proprietary prompt set of a leading AI coding assistant, enabling deep customization, behavioral analysis, and educational insight that isn't available through the standard interface.
All parts of Claude Code's system prompt, 24 builtin tool descriptions, sub agent prompts (Plan/Explore/Task), utility prompts (CLAUDE.md, compact, statusline, magic docs, WebFetch, Bash cmd, security review, agent creation). Updated for each Claude Code version.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Provides over 110 exact prompt strings extracted directly from Claude Code's compiled source, enabling deep behavioral analysis and auditability, as highlighted in the 'Complete Prompt Extraction' feature.
Includes a detailed changelog tracking prompt changes across 142+ versions of Claude Code, allowing users to study Anthropic's prompt engineering strategies over time, as noted in the 'Version Tracking' section.
Serves as the basis for tools like tweakcc, which lets users modify and patch local Claude Code installations with custom prompts, facilitating personalized agent behavior without rebuilding from scratch.
Prompts are categorized into Agent Prompts, System Prompts, Tool Descriptions, and more, making navigation efficient for research or selective modification, as described under 'Structured Organization'.
Prompt extraction lags behind Claude Code releases, and breaking changes from Anthropic can quickly obsolete customizations or require frequent updates, risking instability.
Using tweakcc for modifications requires technical expertise in patching npm or binary installations, and improper changes can lead to conflicts or broken functionality in Claude Code.
Heavily tied to Claude Code's proprietary architecture; insights and customizations have limited transferability to other AI tools, creating dependency on Anthropic's roadmap.
The community around prompt hacking for Claude Code is small, with sparse third-party tooling or documentation beyond this repository, limiting support options.