A large collection of role-based AI agent prompt files (engineering, design, marketing, security and more) that install into tools like Claude Code, Cursor and Gemini CLI.
Qiuner/birdview
A skill for AI coding agents that draws the project's architecture, constraints and planned change scope on a map and waits for confirmation before editing code.
- Stars
- 689
- Forks
- 74
- Open issues
- 0
- Last commit
- 10 days ago
What is it?
Birdview is a skill for AI coding agents that makes the agent show its understanding of a project before it edits anything. It produces a standalone HTML page combining the architecture modules and their relationships, the constraints that apply to the work, and the files and modules the current change is expected to touch. The user reviews this map and confirms the plan before implementation begins.
The tool targets the gap between what an agent did and where its change sits in the system. Logs and diffs show actions and modified lines, but not whether the affected modules or outside-scope neighbors were considered. Birdview records declared task scope, progress and check results in an activity file, and can render an architecture review that compares the current structure with a proposed redesign, with source quotes checked against the repository.
Output is a single HTML file that opens offline in a browser, with light and dark themes and Chinese or English interfaces. Installation uses the skills CLI, and invocation is on demand in Codex or Claude Code, with an optional auto mode. The project is MIT licensed and written in TypeScript, requires Node.js 18 or newer.
Key features
- 01Renders architecture modules, ownership and relationships as an interactive map.
- 02Shows the files and modules a planned change declares, plus directly connected modules outside its scope.
- 03Requires user confirmation of the plan before implementation.
- 04Produces a review page comparing the current architecture with proposed redesigns, with source quotes checked against the repository.
- 05Generates standalone offline HTML with Chinese and English interfaces.
Use cases
- Reviewing an AI agent's planned edits across modules before letting it write code.
- Getting an up-to-date architecture map of an unfamiliar codebase.
- Running an architecture review to compare the current structure with a proposed redesign.
- Checking which neighboring modules a change could affect.
Topics
FAQ
- What is birdview?
- A skill for AI coding agents that draws the project's architecture, constraints and planned change scope on a map and waits for confirmation before editing code. Birdview is a skill for AI coding agents that makes the agent show its understanding of a project before it edits anything. It produces a standalone HTML page combining the architecture modules and their relationships, the constraints that apply to the work, and the files and modules the current change is expected to touch. The user reviews this map and confirms the plan before implementation begins.
- Is birdview free to use? What license does it have?
- birdview is open source under the MIT license. Check the license terms before using it in commercial products.
- Is birdview still maintained?
- Actively maintained. The last commit to birdview was on Oct 9, 2026.
- What can I use birdview for?
- Reviewing an AI agent's planned edits across modules before letting it write code. Getting an up-to-date architecture map of an unfamiliar codebase. Running an architecture review to compare the current structure with a proposed redesign. Checking which neighboring modules a change could affect.
- Where can I get birdview?
- The source code and installation instructions are on GitHub: https://github.com/Qiuner/birdview
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A marketplace of plugins, agents, skills and commands for Claude Code, Codex CLI, Cursor, OpenCode, Copilot and other AI coding tools.
A mobile, web and CLI front end for running Claude Code and Codex sessions from your phone, with end-to-end encrypted sync between devices.
Desktop, mobile, web and CLI app for running and orchestrating multiple coding agents such as Claude Code, Codex and OpenCode in parallel worktrees.
A self-hosted personal AI assistant that runs on your own devices and answers you through chat apps like WhatsApp, Telegram, Slack, and Discord.
A software development methodology and composable skills library that guides coding agents through brainstorming, planning, TDD, and subagent-driven implementation.